What makes an effective research and development robot?

What makes an effective research and development robot?

Mar 18, 2025 | Engineering, InDro Solutions, Industry News, Scott Simmie

By Scott Simmie

 

At InDro Robotics, we sell a lot of robots and drones for the purpose of research and development.

Those devices range all the way from small and highly affordable out-of-the-box solutions like the LIMO PRO right through to highly complex builds for some of the largest technology companies in the world. And that image above? A recent build with a manipulator arm (and many other capabilities) for a client.

But what makes for an effective R&D robot? We put that question to Luke Corbeth, Head of R&D sales. Broadly speaking, he identifies four pillars of research when it comes to R&D. They are:

  • Control
  • Planning
  • Perception
  • Interaction

In fact, Corbeth recently pulled together a graphic explaining these pillars:

R&D Research Pillars Luke Corbeth

CROSSOVER AND CUSTOMIZATION

 

While the four pillars are all distinct, they’re not mutually exclusive. R&D might include both planning and perception, or any other combination of the above. And the focus of the R&D will obviously inform what sensors – even what locomotion – are ultimately required. Does the robot need autonomy, or will the client be coding their own autonomy stack? Is Simultaneous Localisation And Mapping (SLAM) required? Does the robot need to be able to navigate stairs? For our clients, these questions are all discussed in great detail during an initial discovery call with Corbeth.

Sometimes, particularly in lab-based work, an out-of-the-box solution may be all that’s necessary. At Boston University, for example, they have a fleet of LIMOs deployed in the lab for research on multi-agent systems (and other areas). R&D in the field, by contrast, generally calls for a larger and more robust type of robot.

“A project in the lab often means you can get away with a smaller platform,” says Corbeth. “And when people are trying to tackle problems in the real world they’re often using larger platforms.”

While this is generally true, there’s no question some indoor R&D can require incredibly sophisticated robots. One of our more complex builds, which we affectionately named Rosie, is a dual-manipulator robot designed for the Industry 4.0 setting. In fact, the entire lab itself is built for 4IR – with a suite of interconnected devices that share data not only within that location itself, but also with other R&D labs. (It’s actually a super interesting project, which we explored in detail here.)

Regardless, the planned R&D will inform what’s needed – including platform, sensors, etc. That’s where InDro has extensive expertise, not only in integration but in having tested and proven the components themselves. And that saves clients a lot of time and energy.

“If every client or research lab had to build a robot from scratch, it would take them a lot longer to get to the point where they need to be,” explains Corbeth. “And that’s largely why we come into the picture. We help jumpstart these projects and get them to their end goal much faster.”

Because InDro has years of experience building both custom robots for clients and our own products, we’ve learned – sometimes painfully – which components and platforms offer the best value and reliability. And, in conjunction with InDro Forge, we have the expertise for seamless integration,

“We’ve tested a wide range of different hardware and configurations,” he says. “We’ve basically swallowed that pill already so that our clients don’t have to.”

Plus, of course, if a customer already has some components they want to use, such as a pricey LiDAR, we can customise a package to exclude that and save the end-user money.

Below: Rosie, a dual-manipulator mobile robot we built for pick-and-place in a lab doing Industry 4.0 research

 

Rosie

PLUS, OF COURSE, OPEN SOURCE

 

Everything we sell, when it comes to robots, is Open Source and nearly always with ROS 2 (Robot Operating System 2). This is a significant upgrade from ROS 1, which relied on a Master-Slave architecture. By making that architecture more distributed (eliminating that central ROS Master), ROS 2 reduces single points of failure and is more scalable. It’s also what those in the R&D space generally want, as Open Source allows them to easily pull in pre-existing code suitable to their research.

“The main thing with Open Source is not having to start at zero. If everything was Closed Source, you’d have to do everything from scratch,” says Corbeth. “That’s the value of Open Source; you’re building off of the discoveries of your peers, and that dramatically expedites progress for everyone who is Open-Sourcing their projects.”

And a final thing worth mentioning? Support. InDro has built a solid reputation for its after-sale support. From warranties and remote troubleshooting through to site visits (on the rare occasions that becomes necessary), we back what we build. We believe in minimising downtime for our clients so they can get on with R&D.

And remember those four R&D pillars? It’s a great top-level view. But Corbeth also took the time to drill down within those themes to take a far more detailed look at research areas and use-cases. It’s amazing the number of areas where research is taking place (and there are likely even more that could be added):

R&D Research Themes Luke Corbeth

INDRO’S TAKE

 

We’re in kind of a unique position when it comes to helping clients requiring robots or drones for Research and Development – because we’re an R&D company ourselves. In addition to finding the best solution for customers (whether it’s out-of-the-box or a complex custom build), we are continuously developing our own products. Some of those products, like InDro Commander and the forthcoming InDro Cortex have been designed for those clients who want to build or modify their own robots with ease. We truly understand the R&D journey – and have something of a special affinity for clients in that space.

“From the very outset, InDro Robotics was formed as a Research and Development company, so we truly get it,” says Founder and CEO Philip Reece. “It’s in the interests of the entire robotics industry to see advances in the R&D space – so we’re always happy to assist with solutions from the simple to the complex. It really is what we do.”

Want to continue the conversation? Feel free to contact us here. He’s always happy to talk robotics with zero pressure.

Get ready: New RPAS regulations are coming in Canada

Get ready: New RPAS regulations are coming in Canada

Mar 14, 2025 | Industry News, Kate Klassen, Regulatory, Scott Simmie, Transport Canada

By Scott Simmie

 

New Transport Canada RPAS regulations are coming. Precisely how soon is a question of some debate, particularly since the country just swapped Prime Ministers and is in the midst of a pretty unpleasant trade dispute with its largest trading partner.

Nonetheless, the work of government marches on. And in the near future we can expect Canada Gazette 2 to announce significant forthcoming changes to RPAS regulations which will be phased in likely by the fall of 2025.

That means this summer’s flying season, in some respects, will be business as usual. Beyond Visual Line of Sight flights will still require a Special Flight Operations Certificate through Transport Canada. But once the regulations come into force, low-risk BVLOS over sparsely populated areas will no longer require an SFOC, providing the pilot, organization and drone (up to 150 kg) all meet new requirements.

This will open up the door for the industry to carry out long-range BVLOS data acquisition and deliveries in low-risk scenarios without all the paperwork. But there will be some additional barriers to ensure these missions – and the pilots and organizations carrying them out – meet new TC requirements.

With the help of InDro’s Training and Regulatory Specialist Kate Klassen – who’s also a traditional aircraft instructor with multiple ratings – we’ll recap what’s coming, along with how to start preparing for the transition.

 A BALANCING ACT

 

Transport Canada, as regulator, has a delicate task: It needs to ensure airspace safety as much as possible while also allowing the industry to advance and grow economically. That’s what led to the first set of RPAS rules – Part IX of the Canadian Aviation Regulations (CARs) – announced in 2019. Those rules included the requirements for Basic or Advanced RPAS Certificates to operate drones weighing between 250 grams and 25 kilograms. It also laid out the regulations for where VLOS flight could (and couldn’t) take place and made it clear what type of missions would require an SFOC. Prior to Part IX of CARs, every commercial drone flight – including VLOS flight – required an SFOC. So that was a big step forward.

As the industry and its requirements continued to grow, TC started planning for the future. In mid-2023, this issue of Canada Gazette outlined proposed amendments that would ease the path to broader use-cases of drones while adding additional requirements on the safety and planning side of things. It’s a lengthy document, but its broad goals are boiled down to a single paragraph:

“This regulatory proposal would allow operations with a remotely piloted aircraft (RPA) up to 150 kg to be flown within visual line-of-sight and introduce rules for routine beyond visual line-of-sight operations with an RPA up to 150 kg over sparsely populated areas, at low altitudes, and in uncontrolled airspace. The proposal would remove the requirement for a Special Flight Operations Certificate (SFOC) for these operations…”

So the amendments will – in far greater detail – cover two new categories of operation: VLOS operation of drones weighing above 25kg and up to 150 kg (in both controlled and uncontrolled airspace) and BVLOS flights that meet low-risk criteria.

And the flip side of the coin? The paragraph continues: “The amendments include proposed requirements for a new pilot certification, new technical standards for the aircraft and supporting systems, new operational procedures, such as increased distances from airports, heliports, and people, as well as new requirements for individuals and organizations to operate BVLOS…”

It will be a significant step forward for operators, as part of TC’s measured approach to balancing safety and growth.

“This isn’t going to be the set of regulations that opens Canada up to delivering your pizza by drone,” explains Klassen. “But it will allow routine, low-risk BVLOS flights without the need for an SFOC. Detailed planning will still be required, but long-range data acquisition and deliveries in low-risk scenarios will become a lot more common.”

As noted, however, this isn’t a free pass.

“Operators wanting to carry out those low-risk BVLOS missions will need to demonstrate they have the knowledge and skills to do so. Pilots will need to obtain a new Level 1 Complex Pilot Certificate. And while it still covers the same eight knowledge areas as an Advanced RPAS Certificate, it’s to a whole new depth.” says Klassen.

Those with an existing or new Advanced Certificate will be able to carry out:

  • VLOS operations with a medium-sized drone (above 25 kg up to and including 150 kg);
  • Extended VLOS operations (EVLOS), using a visual observer; and
  • Sheltered operations, which would allow the drone to be flown around an obstruction (e.g. a building) without the use of a visual observer.

“You can temporarily put the drone out of your line of sight,” explains Klassen. “Say, fly behind a building or descend below the tree line if you’re comfortable with the safety and managing the risk of those scenarios. A cool example might be firefighters that need to duck the drone behind a plume of smoke so they don’t have to go the long way around.”

 

WHAT CONSTITUTES LOW-RISK BVLOS?

 

This is an important question – particularly when it comes to operating drones up to 150 kg. You wouldn’t want these flying in places that could put people, or traditional air traffic, at undue risk.

The proposed regulations (and we still have to wait to see the final language) indicate that TC will require operators to ensure their missions meet acceptable risk management criteria under SORA, an international guideline which stands for Specific Operations Risk Assessment. This combines calculating the ground risk (people, property, infrastructure) of a specific planned mission (where BVLOS obviously carries a higher risk than VLOS) as well as the air risk – the probability of encountering crewed aircraft in the airspace. The latter is calculated based on the density of traditional aircraft in the proposed operational airspace (higher density equals greater risk), plus any mitigating factors such as detect and avoid systems, robust operating procedures, etc..

The final language isn’t out yet, but the paragraph cited earlier contains the high points. Low-risk BVLOS flights that take place over sparsely populated areas in uncontrolled airspace and at altitudes not exceeding 400′ AGL will not require an SFOC.

There is, obviously, a big difference between VLOS and BVLOS, particularly when we’re talking about larger drones. BVLOS missions will involve more complex planning akin to traditional aviation. If it’s a long-range mission, what’s the anticipated weather on the route 100 km away in two hours? Pilots with their Level 1 Complex Certificate will need to take into account multiple risk factors that don’t generally apply to VLOS flight.

“The knowledge requirements will be a step above Advanced, for sure,” says Klassen. “You’ll be expected to to know about antennas and variables that can impact your C2 link reception, and how to apply that knowledge operationally. You’ll have to be able to look at a proposed route and understand how the environment and terrain features are going to impact your reception range – and even whether or not the proposed mission is even possible.

“We can also expect a lot more as well on crew communications and operating procedures for if things don’t go as planned.”

Below: These graphics, pulled together by Kate Klassen and InDro, cover key aspects of the anticipated new RPAS regulations

new RPAS regs
RPAS regs

WAIT, THERE’S MORE

 

The new Level 1 Complex Certificate comes with a mandatory ground school requirement (you won’t be able to write the online exam without it), as well as a medical sign-off from your doctor to ensure you don’t have any untreated condition that could impact the safety of operations. That declaration from the doctor will be need to be presented at the more extensive in-person Flight Review the new certification will require. TC is also aligning the rules for company ownership to more closely align with existing regs for traditional aviation. These will be more favourable to Canadian citizens and companies with at least 75 per cent Canadian ownership.

Oh, the regs will also permit Extended Visual Line of Sight (EVLOS) flights with additional visual observers, and also allow for brief BVLOS flight if necessary within missions that might have been planned as VLOS only.

There will be much more, including the need for organizations to have an Accountable Executive responsible as the point person for overall ops – what Klassen jokingly refers to as the “one throat to choke” if things go wrong.

The fine print of the final regs, as noted, is yet to be released. But it will be a significant advance for the industry – coming with additional responsibilities and knowledge requirements for those involved. InDro will be ready, with updated courses on its FLYY online drone instruction portal run by Klassen. In fact, there’s already a free prep quiz for the new regs here.

Stay tuned.

 

 

Industry 4.0 and InDro – the evolution continues

Industry 4.0 and InDro – the evolution continues

Mar 11, 2025 | 4IR, Autonomy, Industry 4.0, Industry News, Scott Simmie

By Scott Simmie

 

Many of you will remember the days before smart phones. Same goes for automated tellers, online banking, self check-outs, personal computers, 3D printers – even the internet itself. Technology hasn’t merely marched along; it’s been sprinting at an ever-accelerating pace. What’s more, it’s been doing so pretty much everywhere. From the smart devices that now populate our pockets and homes and vehicles through to autonomous mobile robots in factories, hospitals, warehouses, airports – we are in the midst of an inflection point.

If you’re in the technology industry, this era is known as Industry 4.0. And there’s no question that it is – and will continue to be – utterly transformative.

Let’s take a brief look at how we got here…and where it’s going.

Below: An InDro Robotics Sentinel inspection robot. It carries out complex autonomous inspections before returning to its base to wirelessly recharge

Sentinel enclosure Ottawa Hydro

THE PATH TO 4.0

 

Industry 4.0 is also known by some as 4IR, meaning the Fourth Industrial Revolution. So it’s worth briefly reviewing the other three.

The initial Industrial Revolution began in the UK in the mid-1700s. The development of steam power, water power, and mechanisation paved the path for production of certain commodities at scale. They may seem primitive now, but these were huge innovations at the time. These efficiencies helped vault the UK to a leading economic position and the technology began rapidly spreading elsewhere in the world.

That was followed by three other industrial epochs:

  • The late 1800s, where mass production lines using electrical power marked the outset of the Second Industrial Revolution
  • The late 1960s saw the introduction of computers and other early IT systems, as well as significant advances in automation including simple robotic devices
  • The mid-2010s ushered in Industry 4.0, often described as the integration of cyber and physical systems (more on this in a moment)

To help visualise this, we’ve tapped on Wikimedia Commons, and this graphic from Christoph Roser at AllAboutLean.com

Industry 4.0 Wikimedia Commons Christoph Roser at AllAboutLean.com

THE FOURTH WAVE

 

As we saw, what’s thought of as the Third Industrial Revolution brought computers and early robotics/manufacturing advances onto the scene. Industry 4.0 can be thought of as the logical extension of the third – but with massive technological and data integration advances. As this Forbes article puts it, “The fourth industrial revolution will take what was started in the third with the adoption of computers and automation and enhance it with smart and autonomous systems fueled by data and machine learning…As a result of the support of smart machines that keep getting smarter as they get access to more data, our factories will become more efficient and productive and less wasteful.”

We asked an AI engine for its take, and it came back with a very concise definition: “Industry 4.0 is a term that describes the ongoing technological revolution that is transforming how companies operate, design, produce, and deliver goods and services.”

It also offered, helpfully, the key enabling technologies including: 

  • Artificial Intelligence 
  • The Internet of Things 
  • Big Data Analytics 
  • Augmented Reality 
  • Precision Scanning and digital twins
  • Robotics
  • Advanced manufacturing techniques, including 3D printing

COVID-19, with its extensive isolation and social distancing, played a significant role in companies embracing Industry 4.0. A basic example many can relate to was the growth of UberEats and other food delivery services. The coding and technology – the integration of the cyber and physical words – utterly transformed much of the restaurant industry.

It would be hard to think of a sector that has not been touched by 4IR: Manufacturing, mining, agriculture, pharmaceuticals, aerospace – you name it.

 

INDRO 4.0

 

Industry 4.0 is a massive topic – with implications not only for companies seeking a competitive edge but also for workers. Many companies, according to this excellent McKinsey and Company overview (complete with compelling data and examples of ‘Lighthouses’ – companies at the pinnacle of 4.0), are re-skilling employees hand-in-hand with adopting new 4IR technologies. Europe here has taken the lead over North America.

As for InDro? The company was officially formed in 2014. That happens to be the year generally accepted as the year Industry 4.0 began. And from the beginning, this has been the realm where our R&D has taken place. As a leader in the autonomous robotic space, many of our own inventions and custom builds operate in the Industry 4.0 space. We’re particularly proud of our Sentinel inspection robot (several of which are now working autonomously for a major US energy client), and also Captis – the leading solution in inventory cycle counting and precision scanning for large warehouses and other supply chain assets. InDro Robotics was the technology incubator for Captis, produced by Cypher Robotics. It’s already on the job in Canada, and will soon be deployed in New Zealand.

Below: The Captis cycle-counting and precision scanning system

Sentinel

INDRO’S TAKE

 

Industry 4.0 isn’t just a buzzword. It is a full-fledged transformation leveraging multiple complex technologies working in synergy for greater efficiency. Most of our clients have fully embraced IR4 or are in the midst of that transformation. And we, as always, continue to develop new robots, drones and other products for this new and exciting era.

“Industry 4.0 certainly draws on the framework laid by 3.0, but the technological advances of the past decade have been truly transformative,” says InDro Robotics Founder and CEO Philip Reece. “We are definitely in the midst of a new and exciting era, and InDro will continue to develop intelligent and innovative products for Industry 4.0. And yes, when 5IR eventually comes along…we’ll be ready.”

Want to learn more about how an InDro solution can help your company in IR4? Interested in learning how a private 5G network can offer smart factories a competitive and security edge? Head of R&D Sales Luke Corbeth is always up for a thoughtful conversation.

RIP Romeo Durscher – a true pioneer in the drone field

RIP Romeo Durscher – a true pioneer in the drone field

Mar 5, 2025 | Industry News, Scott Simmie

By Scott Simmie

 

The drone industry has lost a true visionary. Romeo Durscher, who revealed last year he had terminal cancer, passed away peacefully in his sleep March 3.

But while he’s gone, he certainly – truly – will not be forgotten. He leaves an incredible and admirable legacy. His work, and personality, touched countless people in this sector.

As most of you will know, Romeo was an early champion for the use of drones in First Responder work. That passion would ultimately influence the design of several DJI drones and expand use-case scenarios. He worked closely with police departments, tactical agencies, emergency responders and even Interpol over the years. His contributions were recognised with multiple honours.

In fact, in December of 2024, the Law Enforcement Drone Association honoured Durscher with its first-ever Lifetime Achievement Award.

“This prestigious honor is a testament to Romeo’s visionary leadership and extraordinary contributions to public safety through the adoption of drone technology,” said LEDA in its announcement.

“Long before drones became integral tools for saving lives and protecting communities, Romeo foresaw their transformative potential. With unwavering passion and dedication, he turned that vision into reality, revolutionizing public safety operations around the world.”

And he will be very much missed.

“He had a way of lighting up a room, of making everyone feel seen and valued,” wrote Airwards Founder Richard Nichols in a post on LinkedIn. “His passion for drones and the power of their potential to do good was infectious, and he inspired countless people to pursue their own dreams in this industry. He leaves behind a legacy of innovation, kindness, and genuine human connection.”

Below: Romeo receiving an award on behalf of DJI at the 2019 Aerial Evolution Association of Canada conference

Romeo Durscher

THE ROAD TO RPAS

 

Romeo obviously didn’t start out in the drone world. RPAS as we now know them, didn’t exist when the Swiss-born (and multilingual) Durscher started out in the world of banking. He began as an intern and it wasn’t long before he was promoted to the position of Investment Banker. A few jobs later (including working as a rock promoter, even organising tours) and he landed at Stanford University. He spent nearly 13 years there, where his interest in technology had a chance to flourish.

Romeo was Stanford’s Senior Manager on its Solar Dynamics Observatory – a partnership with NASA. He often spoke fondly of those years, and would later tell DroneLife it was there that he first built a multi-rotor.

“In 2010, when I was still deeply involved in NASA missions, Mark Johnson and I began experimenting with a new kind of technology. Multi-rotors were practically unheard of, but we found components from all over and pieced together our own drones. I vividly remember the first time we flew one. For Mark, it was clear this could be a powerful tool for his forensic analysis business. For me, it unlocked something new—a fresh perspective from the air,” he told the website.

“Something about this new technology grabbed me. I knew it had a future beyond NASA, especially in public safety. The memory of my Swiss military service, wishing for a real-time aerial view, fueled my desire to bring drones to new spaces.”

 

DJI

 

Drones were still in their very early stages, but Durscher could see the potential. This was the space he wanted to pursue. He joined DJI in January of 2015 as Director of Education, back in the days when a GoPro attached to the first-generation Phantom was seen as pretty much state-of-the-art.

As the technology advanced, so too did use-case scenarios. And Durscher was at the forefront, becoming the Senior Director of Public Safety Integration. That job entailed working closely with fire rescue, EMS, law enforcement, tactical agencies – you name it. And not only did he help expand their capabilities, he also spent endless hours listening to those First Responders. What other use-cases might drones be helpful for? What other features – whether hardware or software – might make these tools even more useful?

He took that feedback directly back to the engineering teams at DJI in Shenzhen. That intel directly influenced the final version of many of DJI’s Enterprise series of drones. The Mavic 2 Enterprise, with its options of thermal imaging, a loudspeaker, and a bright light, owes much of its heritage to Romeo Durscher. He also played a role in the design of the M200, M300 and even the third-party Epson Moverio AR glasses.

But that’s not all. Durscher also built a reputation for getting down in the trenches when necessary. In November of 2018, the devastating Camp Fire broke out in California. Within 12 hours, that wildfire had consumed 55,000 acres and thousands of homes. More than 100 people lost their lives.

Durscher and DJI immediately jumped in, working with multiple agencies in the most ambitious high-resolution drone mapping project in history at that time. The entire devastation was mapped – providing an essential tool for First Responders, urban planners – and even residents who wanted to see what had happened to their neighbourhood. It was an absolutely immense task, and Durscher was a key lynchpin for the overall organization and implementation.

And yet, in true Romeo fashion, he downplayed his own role – instead extending credit to all of those involved and the technology itself. Those who knew Romeo will recognise this as a character trait.

Below: The massive map database of the Camp Fire

AFTER DJI

 

Romeo left DJI in December of 2020 with an already incredible legacy. But he wasn’t done yet. From there it was off to Auterion, where he became a champion of Open Source and broader robotics integration as Senior Vice President of Strategy. He also served as a Director at FLYMOTION, worked hands-on as an operator and advisor with Florida’s Pasco Sheriff’s Office, and spent several years as a Strategic Advisor with UTAC, which carries out rigorous drone and ground robotics training events for First Responders at its purpose-built facility.

Plus, of course, there was rarely a conference or industry gathering where Durscher wasn’t a sought-after speaker. He was an absolutely brilliant communicator, breaking down even highly complex processes or concepts into easy to digest narratives. He was, in addition to everything else, a born storyteller.

He was also, to those lucky enough to know him, a friend. InDro’s Scott Simmie, who knew Romeo for many years, reconnected at the AUVSI conference in Florida, where the picture below was taken. It’s followed by an extended interview carried out when Scott was running DroneDJ

Romeo

ROMEO’S TAKE

 

We’ve tried to touch on the highlights, but it’s impossible to capture them all. Anyone in the drone world knows that Romeo’s contributions were immense. But when it came to Romeo’s own assessment of his life, he didn’t focus on career accomplishments as the thing that counted most. It was his connections, his friendships.

When Romeo learned he had terminal cancer, he set up this website. He could have used it to list his many career highlights. He didn’t. It was, and is, rather a celebration of a life well-lived, and of lessons learned. In that spirit, we leave the final words here to Romeo himself.

“I suppose we all have wondered how we would react and what we would do if ever confronted with this sort of challenge,” he wrote. “I’m hoping that by sharing my own journey I can help others understand and appreciate how important it is to live passionately and freely for the moment. Indeed, now so many of even the smallest of incidents and minor pleasures throughout my life come back to me as welcome memories and warm reminders of all the wonderful lives that have intersected with my own….

“I’ve been blessed with so many friends who have cherished me and helped me attain so many goals. I believe and hope I can convince you that pursuing your dreams should be a constant motivation throughout your life. That pursuit — and the experiences you have and people you meet in the process — will give you great gifts that are more precious than any material gains. Of all the things in this world that I’m thankful for, my friendships are at the very top of the list. True wealth is measured by friendship – not money.”

Amen. And Godspeed, Romeo.

I, Robot: The Humanoids are here

I, Robot: The Humanoids are here

Feb 27, 2025 | Autonomy, Humanoids, Industry News, Robotics, Scott Simmie

By Scott Simmie

 

You might own a robot without even realising it.

Have a Roomba? That’s a robot. And a drone? That’s a flying robot. Even a Tesla, in Full Self-Driving mode, is a robot.

There are a lot of definitions out there – but one we particularly like comes from Maja Matarić, a computer scientist, roboticist and AI researcher at the University of California. In her book, The Robotics Primer, she concisely defines a robot as “an autonomous system which exists in the physical world, can sense its environment, and can act on it to achieve some goals.”

Whether that goal is to vacuum your floor, capture aerial data, or weld a part in a factory, we feel this is a really clear definition. It also doesn’t delineate between platforms: A robot that fits this bill could be stationary, wheeled, a quadruped or even a humanoid.

And it’s that last platform – humanoid – that’s been getting a lot of buzz recently. Numerous companies are now manufacturing robots that resemble human beings in their form factor. And, as it turns out, for very good reasons.

Below: Ameca, a robot built by the UK’s Engineered Arts, is known for its eery ability to mimic human expressions

Ameca robotics AI

WHY HUMANOID?

 

The idea of a humanoid robot has been around for longer than you might think. Leonardo Da Vinci designed – and possibly built – an automaton in the late 15th Century. It’s known currently as Leonardo’s Robot or Leonardo’s Mechanical Knight. According to Wikipedia, “The robot’s design largely consists of a series of pulleys that allow it to mimic human motions. Operational versions of the robot have been reconstructed by multiple researchers after the discovery of Leonardo’s sketches in the 1950s.”

It appears that the purpose of this design was for entertainment (which also fits the definition of a goal), but it fell short when it comes to sensing its environment and autonomy. Still, it’s fascinating to know the Italian inventor turned his attention to designing a mechanical device in human form way back then.

It would take another half a millennia before the first true humanoid robot would be built. In the early 1970s, the Wabot was unveiled in Japan. It was anthropomorphic, with two arms and two legs. It also contained a vision system, audio sensors and could speak in Japanese. According to this overview, “It was estimated that the WABOT-1 has the mental faculty of a one-and-half-year-old child.”

Below: A modern reproduction, based on Leonardo Da Vinci’s sketches, of his “Mechanical Knight” complete with inner mechanisms. It’s followed by an image of Wabot-1 from 1973

Da Vinci humanoid robot
1973 Wabot humanoid

THE HUMAN ADVANTAGE

 

Why create a humanoid in the first place?

Well, there are certain advantages to a human form factor, particularly when it comes to carrying out repetitive tasks in the real world. And the reason? The world around us has been built for humans. If there’s an existing task carried out by people, say pick-and-place, the infrastructure for that task has been created with humans in mind. That means conveyor belts, shelving, cupboards etc. are all designed for the average human. If you build a robot in a human-like form and roughly to scale, that’s a big advantage.

“You don’t need to change the surrounding infrastructure to accommodate the robot,” explains Head of R&D Sales Luke Corbeth.

“The end result obviously is faster deployment. This applies to factories, homes, hospitals, pretty much any use-case. None of these locations need to be robot native to effectively leverage a humanoid robot because they’ve been built for people.”

In fact, humanoid robots have already been deployed on some factory floors. They’re ideally suited to repetitive tasks such as picking up an item and moving it from one location to another – and contain tactile feedback sensors in their manipulators to calculate appropriate grip strength. They could also be deployed, says Corbeth, in environments built for humans – but which may pose hazards. An example, says Corbeth, might be for inspections or maintenance inside a nuclear facility in a radioactive environment.

“There are a lot of dexterous tasks people are doing today that are very challenging to automate because they require high levels of precision,” he says. “These are perfect tasks for humanoids.”

Looking down the road, many foresee an era when humanoids are affordable enough – and capable enough – for deployment in homes. There, they could carry out some of the more mundane household tasks like cleaning or clothes washing, perhaps even elder care and companionship.

A growing number of companies are now in the humanoid space, including Tesla (Optimus), Agility (Digit), Boston Dynamics (Atlas), and Figure (Figure 02). InDro Robotics is a distributor for Unitree, and carries the G1 humanoid and H1 and H1-2 research and development models. (We can also modify these robots for specific use-cases.)

The base version of the G1 sells for $21,600 US – which is surprisingly reasonable for a humanoid form factor. Corbeth says the current offerings are a result of a “perfect storm” across multiple advances in AI compute, battery, sensor and manufacturing technologies. The more advanced H1 sells for $99,600 US and is better suited for complex R&D.

 

WHAT’S NEXT

 

Humanoids are already in the real world. With further and inevitable advances in AI, Machine Vision and Machine Learning (as well as sensors, manipulators, etc.) it’s safe to assume that humanoids will only get smarter and better at smoothly carrying out fully autonomous tasks.

“I think that it will be probably, realistically, three to five years before you see walking humanoid robots around people all the time,” Dr, David Hanson, Founder of Hanson Robotics recently told the South China Morning Post.

“I think we are entering the age of living intelligent machines. It’s coming. Machine consciousness, self-determining machines…it’s on its way. And if we see that happen, then we want to make sure that we make the AI good, compassionate, able to connect and want the best for humans.” Yes, indeed.

And a final note. At some point, these humanoids will be good enough to manufacture themselves. That’s historically been something in the realm of science fiction. However, a recent TechCrunch story pointed out a new partnership between humanoid developers Apptronik and manufacturer Jabril.

“This means that should everything go according to plan, the humanoid robot will eventually be put to work building itself,” says the article.

Below: A C-NET video outlines developments expected in this field in 2025

INDRO’S TAKE

 

Because we sell and modify humanoids in addition to designing and building our own robots (and robots for clients), we’re obviously interested in this space. While we don’t have plans to develop our own humanoid (yet), we are currently working with the Unitree G1 and H1 models to evaluate and enhance their capabilities. And yes, we’ve already sold these to customers.

“Humanoids are a logical progression in robotics,” says InDro Robotics Founder and CEO Philip Reece. “While they’re not the solution for every use-case, they have a clear role in carrying out repetitive or even dangerous tasks that are currently carried out by humans. I suspect, in the not-so-distant future, humanoids will be working alongside people in an ever-increasing number of settings.”

Interested in learning more? Contact us here.

Sense, solve, go: Does Waymo herald the future of autonomous vehicles?

Sense, solve, go: Does Waymo herald the future of autonomous vehicles?

Feb 21, 2025 | Autonomy, Industry News, Scott Simmie

By Scott Simmie

 

During a recent trip to California, I had the opportunity to ride in a Waymo.

I’d certainly read about Alphabet’s autonomous car-for-hire service in the past and work for a company that builds robots and autonomy software. So it seemed a natural, while in San Francisco, to download the app (similar to Uber) and hail an autonomous vehicle.

Within a couple of minutes, a Waymo vehicle arrived at the pickup point – just a short walk from where it had been summoned. It pulled up, LiDARs spinning, waiting for me to climb in. I put a hand on the door; it was locked. A quick glance at the app and I saw an “unlock” feature. Then I was inside.

And then, with some ambient music playing in the background (you have the option to turn it off or select something else), we – meaning the car and I – were off. A display showed a digital representation of what the vehicle was seeing in its surroundings, including parked vehicles and pedestrians. The electric Jaguar quietly accelerated to the speed limit, obeyed all traffic rules, and smoothly adjusted for unexpected occurrences. When the driver of a parked vehicle opened the door to exit, the Waymo liquidly arced a safe distance away. With the steering wheel making smooth turns and constant smaller fine adjustments, it was like being in a vehicle with an invisible, silent driver at the helm.

I had full faith in the technology and actually preferred it to a standard rideshare. Waymo’s safety data (which we’ll explore later) had reassured me the drive was going to be statistically safer than riding with a human driver. Plus, there was no need to engage in small talk. When the ride was done, I simply exited without being prompted for a tip.

I’d been aware of Waymo since it first deployed. I also have a friend with a Tesla who has the Full Self-Driving package. He commutes twice a week from well outside Toronto into the GTA without any inputs beyond setting his destination. He foresees a day, long promised by Elon Musk, when his own vehicle will earn him money by working during off hours as an autonomous taxi.

The technology for fully autonomous vehicles is basically here – arriving both sooner and later than some had predicted. But what does that mean for the future?

Below: A Waymo Driver waits patiently at an intersection – while another Waymo Driver glides past. Photos by Scott Simmie 

SENSE, SOLVE, GO

 

Owned by Google parent company Alphabet, Waymo states “We’re on a mission to be the world’s most trusted driver. Making it safer, more accessible, and more sustainable to get around — without the need for anyone in the driver’s seat.” It calls its service, the app and the car together, Waymo One. The system, the hardware and software, are referred to collectively as Waymo Driver.

Commercial rollout began in Phoenix, Arizona in 2020, with testing in San Francisco commencing late the following year. It’s now also operating in Los Angeles, with expansion into Atlanta, Austin and Miami next. It uses a Jaguar I-PACE electric vehicle as its base, heavily outfitted with an array of sensors. We’re talking a lot of sensors.

In total there are 29 cameras, six radar and five LiDAR units (including a roof-mounted 360° LiDAR). These sensors allow Waymo Driver to fully capture its environment up to three football fields away. Powerful AI, machine vision and machine learning software continuously crunch predictive algorithms allowing Waymo to understand where a pedestrian, cyclist or other vehicle will most likely continue based on current trajectory. And, of course, if one of those moving subjects suddenly does something unexpected/unpredicted, the system quickly readjusts. Waymo touts safety stats (which we’ll explore later) it says proves Waymo Driver is far safer than a human driver.

Waymo boils down the entire process to three words: Sense, solve, go. Waymo Driver senses its environment using the sensors mentioned above (it also has an array of External Audio Receivers – EARs – which alert the system if they detect sirens, etc. It can even echolocate the location of said sirens and will understand if it needs to pull over). The algorithms solve the challenge of safely moving through that environment, including moving objects, and then it’s go time. Those three simple words represent a decade and a half of intense R&D, development of its own sensors, and a huge capital expenditure.

Google first began exploring self-driving vehicles in earnest back in January of 2009. By the time it revealed this publicly, it had already done extensive R&D and testing. But it wasn’t until the fall of 2015 that the first solo member of the public climbed into a Waymo in Austin, Texas and took the vehicle for a ride on city streets. That passenger was Steve Mahan, who is legally blind. It was the first time in 12 years that he’d been alone in a car. It would be another five years before the first rollout to the public.

During that five years, both the car and sensor package – along with the software – evolved considerably. Just compare videos below; the first shows Steve Mahan on that historic trip in 2015, the second is an updated video explaining the fifth-generation Waymo Driver.

SAFETY FIRST

 

 

Waymo could pitch its offering on a number of grounds: Sustainability, convenience, cool factor. But instead, it focuses its customer-facing marketing on safety. Waymo Driver, it says repeatedly, is far safer than a human driver. It has proven that in many millions of miles on the streets, it says, and billions more in simulation.

The statistics Waymo publishes are based on Incidents Per Million Miles (IPMM) of driving – and compare its own rates of incidents with a benchmark of IPMM involving human drivers. Whether it’s airbag deployments, crashes with a reported injury, or incidents where police are notified, Waymo’s stats are consistently a fraction of those involving people at the wheel. In more than 33 million miles of driving, Waymo touts these as the results:

  • 81 per cent fewer airbag deployment crashes
  • 78 per cent fewer injury-causing crashes
  • 62 per cent fewer police-reported crashes

That’s clearly a significant reduction, and to most people would indicate that Waymo is safer than taking a ride with a stranger (or even friend) at the wheel. The statistics include accidents where other drivers were at fault, but does not separate them out – so we can’t actually see what percentage involved an error from the Waymo side. Waymo has previously stated that the majority of these incidents were the fault of human drivers, and that there have been but two accidents involving injuries where it expects to pay out insurance liability claims.

But even rare incidents can quickly become high-profile. In Phoenix, an empty Waymo that had been summoned by a customer crashed at low speed into a telephone pole in an alley. Had a human been at the wheel, we would never have heard of it. But because it was a Waymo, the incident led newscasts. Why is that? Well, we expect perfection in systems like these. And that seems a reasonable expectation if you’re going to trust your personal safety to a driverless car. It just can’t make mistakes. And that’s why Waymo quickly issued a recall for a software fix.

“This is our second voluntary recall,” Katherine Barna, a Waymo spokesperson, told TechCrunch. “This reflects how seriously we take our responsibility to safely deploy our technology and to transparently communicate with the public.”

In May of 2024, the US National Highway Transportation Safety Agency (NHTSA) informed Waymo it was investigating 22 incidents involving its vehicles (and subsequently added an additional nine incidents), stretching back to 2001. Many of those incidents were described by Forbes as “surprisingly minor” and 11 of those incidents were culled by the NHTSA from social media reports of the vehicles driving in an unusual fashion (such as using the oncoming lane to avoid traffic problems). The most serious was the aforementioned pole collision.

We were unable to find any reports of Waymo incidents involving a serious injury. The one fatality involving a Waymo occurred in January of 2025, when an unoccupied stationary Waymo stopped at a traffic light was one of several vehicles hit by a speeding car. One person and a dog died in that incident, but because a Waymo was tangentially involved it made the headlines. It is the only case we can find involving a fully driverless vehicle where a fatality was involved – and in this case the vehicle was completely passive. (There was a pedestrian fatality in 2018 involving an autonomous Uber vehicle. In that incident, which occurred in Tempe, Arizona, a human safety driver was in the driver’s seat. She was watching television on her phone when the accident occurred and subsequently pleaded guilty to endangerment.)

While Waymo has an excellent track record, there have been incidents. But with each incident where Waymo Driver has somehow made the wrong decision, it’s reasonable to assume it was followed by a software fix. And here’s where a fleet of autonomous vehicles have a definite advantage over people: That tweak can be instantly applied to the entire fleet.

Still, there are skeptics who argue that – despite those millions of miles of driverless passenger trips – Waymo does not have enough data upon which to draw sound conclusions.

“We don’t know a lot. We know what Waymo tells us,” Philip Koopman, an expert on autonomous vehicle safety at Carnegie Mellon University, told the Miami Herald. “Basically you are trusting Waymo to do the right thing.”

 

THE FUTURE

 

Autonomy is hard – and it takes time: Google and Alphabet have invested more than 15 years of continuous engineering for Waymo Driver to reach this level of technological maturity.

Now, Waymo is rolling out to more cities. Remember those 672 Jaguars that had the software upgrade? They’re just a fraction of the 20,000 I-Pace vehicles Waymo signed a contract with Jaguar to purchase. Plus, the company recently announced that its six-generation vehicle – a Chinese-made electric minivan – is next up for testing and deployment. From all external appearances, Waymo shows no sign of stopping (except at red lights, of course). In 2024, it carried out four million autonomous rides – four times more than its total of trips over the previous four years. Rides in 2024 tripled to 150,000 per week. And the company calculates “Waymo riders helped avoid over 6 million kilograms of CO2 emissions.”

That’s all great. But for any commercial enterprise, even if it’s willing to absorb costs during rollout, the ultimate test will be the bottom line. Will Waymo prove profitable?

We can’t say for certain – and Waymo’s current financials are somewhat invisible to the public, as they’re bundled in with several other projects Alphabet projects. But some analysts predict Waymo, the clear leader in autonomous rideshare, will ultimately win a significant piece of the market. An analysis on Nasdaq.com predicts Waymo could prove over time to be the jewel in Alphabet’s crown.

“Uber does more than 200 million rides each week,” states the story. “Let’s let that sink in. So if autonomous rides can capture even half that market, that would mean 100 million rides per week…If Waymo can capture about one-third of the $1 trillion autonomous rides market, it could generate annual revenues of around $300 billion.” Enough, suggests the story, to double Alphabet’s stock price.

That’s a big prize. And, clearly, incentive for Alphabet and Waymo to continue on the road to profitability.

 

Below: The LCD display for rear Waymo passengers. Note the option to “pull over” if you unexpectedly need to end your ride early

INDRO’S TAKE

 

Because we’re deeply involved in the autonomous space, we obviously take great interest in Waymo and other deployments of autonomous technologies at scale. Waymo Driver is different from most other applications, though, because it’s transporting human beings. There is very little – if any – room for error.

“We can’t predict the future, but – like algorithms – can make informed predictions with available data,” observes InDro Robotics Founder and CEO Philip Reece. “Waymo appears to be heavily invested in continuously making a good safety record even better – and has the engineering and financial resources to do so. I suspect Waymo, and its competitors, are here to stay.”

For more on Waymo, check out its website. And, if you’re in one of the growing number of cities where it operates, download the app and let Waymo Driver take the wheel.