AI, Level 5 Autonomous Driving

Autonomous Driving is not one Disruptive Market – It’s a Disruption to our Way of Life

tony zarola

Tony Zarola
General Manager at Analog Devices

we.CONECT had the chance to meet with Tony Zarola, General Manager at Analog Devices to talk about the technical challenges of automotive AI and the role cognitive computing plays in self-driving car technologies.

Tony Zarola, General Manager of Inertial Sensors, is a 30 year veteran of Analog Devices and is based in ADI’s Wilmington, MA facility. Prior to managing the inertial sensor business, Tony was the Marketing Director for Vital Sign Monitoring Technology in the Healthcare Business Unit and led the initiative that saw ADI become a leader in remote health monitoring. In his current role, Tony is building an inertial sensor technology franchise to support the emerging autonomous systems market and developing solutions for navigation assistance, platform stability as well as system health monitoring.

we.CONECT: What are your main responsibilities in your current role?

Tony Zarola: I am the General Manger for the Inertial Sensors Technology Group at Analog Devices. My main responsibilities are managing the developing of MEMS products that are sold in to the broad markets that ADI serves (Automotive, Healthcare, Industrial and Consumer mainly) and managing business development targeted towards solutions for Navigation and Platform Stability and Vehicle Health Monitoring in the Autonomous Systems market – including Automotive (Private and Commercial vehicles), Smart Agriculture, Drones, Construction etc.

we.CONECT: What fascinates you most about autonomous driving?

Tony Zarola: The breadth of technical challenges that hundreds of organizations across the globe are attempting to solve. New frontiers in engineering are being created that are driving technological advancements in many areas including  a vast array of sensing modalities (e.g. Imaging Radar, Solid State Lidar and High Performance Inertial Measurement Units), advanced communications (e.g. V2X, 5G, Ethernet) to artificial intelligence (e.g. machine learning, neural networks).

That’s just one element of the vast amount of exciting activity that Autonomous Driving is creating others include new business models, a restructuring of the automotive industry with new entrants to the market, consideration to how infrastructure will play a role in the autonomous roll out. The list is pretty much endless, that it’s hard to say the Autonomous driving is one disruptive market – it’s a disruption to our way of life (in a positive way!).

"It’s hard to say the Autonomous driving is one disruptive market – it’s a disruption to our way of life."

we.CONECT: What are your predictions for autonomous driving?

Tony Zarola: Couple of thoughts here.. First I think there are two areas where fully Autonomous vehicles will be deployed : Firstly, L4 Geo-Fenced Ride share and Commercial Delivery and Service Vehicles (e.g. Delivery robots, AGV’s, Geo-Fenced  trucking along defined routes).
Secindly, with the involvement of city planners and highway authorities I believe that the necessary controls and regulations will be in place that will enable these two scenarios to become a reality. The economic benefits that these autonomous will yield through the efficient 24/7 movement of people and goods will open new commercial opportunities for both private businesses and city councils and municipalities.

My other prediction is that the L3 market will evolve over the next few years and that the evolution of automatic driver assistance will force new thinking around how drivers interact with the vehicle and vice versa. The risk with L3 is the creation of a complacency with the auto-pilot features that give rise to drivers not being prepared to re-engage with the vehicle at critical moments. While this will likely put a constraint on the L3 vehicle use cases, it will give rise to new technologies that will be developed that will attempt to counter the risks (.e.g Driver condition monitoring including  health and activity).

we.CONECT: What role does AI and cognitive computing play in self-driving car technologies?

Tony Zarola: AI will play a huge role in self driving cars whereas I believe (from the layman’s perspective) that cognitive computing is some way down the road. With so many established as well as start up companies driving the investment of hundreds of millions of dollars in AI, it can’t not play a major role. This is why I believe that the combination of AI and intelligent high performance sensors will dominate for many years before true machine cognitive computing will become a reality.  We saw this week alone how much further AI has to go before all corner cases of autonomous driving are addressed (and we still don’t know what all the corner cases are). So heavily reliance on high performance sensors such as Radar, Lidar, IMU’s and vision are still critical to go along with AI.

we.CONECT: There is a lot of confusion around autonomous driving. What are the different routes to level 5 and what are the biggest challenges to get there?

Tony Zarola: One of many of the major challenges to getting to true L5 autonomous cars in the non-homogenous situation we will be in for the next 10 to 20 years at lease. Having both autonomous and non-autonomous vehicles on the road (not to mention cyclists, pedestrians etc..) will create numerous scenarios that with just one or two catastrophic events will put a dent in to the roll out of L5 cars. Public acceptance of autonomous cars is still not there today – but public acceptance of any new disruptive technology has always been fraught.
One way to get to L5 I think would be through the evolution of L3 vehicles. Through a gradual evolution of the L3 auto pilot features where over time the driver is offered less time in control of the vehicle to the point where the car is pretty much at L5. However the L3 challenges mentioned earlier will need to be addressed.

we.CONECT: How is your company developing deep learning capabilities? What are the challenges?

Tony Zarola: ADI’s machine learning developments are in the context of the sensor technologies we develop (Radar, Lidar and Inertial Sensors for instance). We are exploring the use of machine learning to improve the performance of the sensors in the autonomous systems.

we.CONECT: Thank you very much for taking part in our interview.

About Analog Devices:
Analog Devices defines innovation and excellence in signal processing. ADI’s analog, mixed-signal, and digital signal processing (DSP) integrated circuits (IC) play a fundamental role in converting, conditioning, and processing real-world phenomena such as light, sound, temperature, motion, and pressure into electrical signals to be used in a wide array of electronic equipment.

Leave a Reply

Your email address will not be published. Required fields are marked *