Autonomy Answers — Meet Keith Leung
This ongoing series features questions and answers from professionals from both inside and outside of Trimble. Our goal is to glean and share their insights, experiences and opinions about autonomy and its range of applications across a variety of industries.
Keith Leung is an engineering manager with Trimble Autonomy, currently leading an R&D group focused on enhancing autonomous solutions with advanced robotic navigation technologies. Having joined Trimble in 2016, he most recently created the Trimble Autonomy simultaneous localization and mapping (SLAM) framework. He received his Ph.D. from the University of Toronto after earning his graduate degree in mobile robotics and SLAM, and has authored 30 scholarly papers on robotics navigation.
Meet Keith Leung, engineering manager, Trimble Autonomy
Q: How do you foresee the impact of autonomous solutions / technologies on consumers’ lives?
"From the consumer perspective, I think the changes that will feel most impactful would come from the increase in mobility as self-driving technology becomes more readily available. In car-centric societies (e.g., North America), there will be a change in how we perceive and use transportation. We will go from serving as operators to our own vehicles to being riders on-demand. Some of these technologies will funnel into new consumer robot products that enter our homes. Overall, there will be a perceived improvement in our standard of living. I think the larger societal impact would come from the increased efficiency of industrial machines leading to higher yield and productivity, and more reliable movement of goods. Particularly in agriculture, I hope this leads to better distribution of food availability and sustainable food costs.”
Q: In your opinion, how will people’s lives be enhanced by the continued progression of driverless technologies?
"In the shorter term, I believe driverless technologies will provide a higher level of situational awareness and safety for drivers — not only for those commuting to work, but also for operators of industrial vehicles. Existing sensing technologies already enable passive systems that inform drivers of lane departure, or provide blindspot warning. Today, a growing number of road vehicles are capable of lane following, and automatic braking. Newer technologies allow all dynamic objects in a vehicle’s vicinity to be tracked, and we have our own version of this technology in house. It enables better decision making by a driver, or a computer. Going forward, autonomous technologies will reduce the cognitive load for vehicle operators. In industrial vehicles, operators can divert their attention from mundane driving work to more productive decision making tasks. Using the SAE standard, the above covers the capabilities up to level-3. Level-4 and above is where we will begin to see enhanced mobility for people.”
Q: What do you most look forward to seeing become reality?
"This first part of my answer might sound a bit out there — human civilization adapting to a post-scarcity economy. While I do not see this happening in my lifetime, I do wish for it to become a reality. Beyond my dreams for the future, I would like to see level-5 autonomous cars on roads that are accessible by the general public. To put us one step closer, I’d like to be able to buy lidars at the same price as radars.”
Q: What’s been your greatest professional accomplishment? What work are you most proud of?
"When I joined Trimble, I was privileged to be given the freedom to innovate and build what I felt to be useful in our journey for autonomy. I focused on creating our own lidar-based simultaneous localization and mapping (SLAM) framework, allowing an autonomous vehicle to locate itself and model its surrounding environment, in the absence of GNSS. Lidar SLAM is not unique to Trimble, but I believe we possess uniqueness in underlying design philosophies. Some of the concepts in our framework have only recently appeared in academic publications. When an early prototype was established, we deployed it in operating environments such as underground mining tunnels, fruit orchards, vineyards, construction sites, industrial manufacturing facilities and rail yards. The system worked in every single application. Our most ambitious deployment was on the streets of Las Vegas where we gave live demos during the Trimble Dimensions user conference. These successes led to commercialization of the technology and our current map-based localization product, which integrates with Trimble inertial navigation systems to provide high-availability positioning for users. I credit my team for taking the initial framework to where it is today.”
Q: Can you share a ‘breakthrough’ or ‘a-ha’ moment in your own journey of implementation of autonomous capabilities?
"One particular deployment we conducted recently was pivotal in our SLAM framework’s development history. Upon arriving at our client’s test site, we were able to set up, demonstrate map-based localization, and pack up within the span of a day. At the end of the demonstration our client challenged us with a scenario that is outside the normal operation limits of our system. Meanwhile, their engineers literally stood watching behind my (sweaty) back. Ultimately, our system performed very well, and the applause from our client engineers made this deployment really feel like a victory. More importantly, it made me realize that our system is now “dialed in.” From that point forth, deployments changed from opportunistic try-and-see exercises to confident engagements where we focused on promoting our technical competency.”
Q: What would you tell someone who’s eager to work on autonomous solutions? What do they need to know, have within them or be excited about to do this work?
"Myself, I took the academic route — from engineering to graduate studies in a specific area of autonomy and going to a post-doc researcher before coming to Trimble. For those following this path, hands-on experience in machine learning would be advantageous. Software engineering proficiency is also expected. However, I would consider technical domain expertise as part of what is necessary to create an autonomous solution. A robot that performs the fundamental functions of sense, perceive, plan, control does not provide value unless those capabilities are incorporated into a workflow that is useful to our partners and customers. We need application domain experts to create and optimize those workflows. We need application engineers to deploy and field our solutions and obtain feedback from the end users. We need product management and business development. The fun part in all of this is problem solving. The excitement of seeing a fully functioning autonomous system is short lived. Luckily for us, or at least the engineers, there will always be new corner cases to curse at and solve.”
Q: Is there a favorite toy/game/activity from your childhood that steered you in the direction of working in the autonomous tech field?
"When I was a few years old, my dad bought a remote control toy boat. We lived within Hong Kong’s urban jungle. Luckily, there was an artificial pond underneath a busy highway bridge where we could go play. I recall being fascinated by how the propellers and the rudder activated from the remote. I think that was my first exposure to a vehicle controlled without an onboard operator. Growing up, I also watched my fair share of mecha / robot anime series. At a young age, there is no distinction between a sci-fi fantasy robot versus a real one, but it doesn’t matter because all robots are awesome. I can’t say for sure if these long-ago experiences steered me into engineering and in pursuing graduate degrees in robotics, but I still love making and seeing robots go.”
Q: In your opinion, what’s the closest thing to real magic that you have experienced in your life so far?
"One of Arthur C Clarke’s Laws states that “Any sufficiently advanced technology is indistinguishable from magic.” For me, the creation of life and consciousness is one of these advanced sciences that we have yet to fully comprehend. Witnessing the beginning of a new life, and hearing the first cries of my children when they were born is the closest thing to magic I have experienced.”