Are VRU Sensing And AI The Keys To Accelerating Autonomous Driving?

By Mark Mitchell, Head of Research, Covenant Venture Capital

Distracted driving is one of the leading causes of car accidents according to the 2020 National Highway Traffic Safety Administration (NHTSA) Traffic Safety Facts Annual Report. While drunk driving and speeding are the top two reasons for fatal car accidents (combined 32%), a distracted driver is estimated to be the cause of 25-50% of motor vehicle accidents, with 5-10% being fatal. When viewing the NHTSA table on related factors for drivers involved in fatal crashes, almost 77% of crashes can be attributed to human error. Indeed, a controversial 2016 study by the NHTSA attributed over 90% of crashes to human error.

While we may automatically assume distracted driving means texting while driving, 41% of major car accidents are the result of recognition error which includes poor attention and distraction. 33% of accidents are the result of decision errors, which include misinterpreting other drivers’ moves, and misreading the speed of turns. This means that almost 3 in 4 automobile accidents are due to the driver making a miscalculation. And when we add smartphone penetration in the U.S. growing from 22% in 2008 to 82% in 2023, and 92% of U.S. drivers using their phones behind the wheel to the equation, distracted driving (and human error) is arguably a national epidemic. 

With the rapid advancement in automobile and cellular technology, we are approaching an age where the human decision making element in driving can be eliminated or reduced to make incremental improvements to automobile safety. Indeed, cellular technology and AI will be critical for autonomous driving as the near real time response capabilities allow for lidar, GPS, and other high latency technologies to be redundancies versus primary proximity systems. 

One of the companies that may seemingly advance autonomous driving while trying to solve the problem of Vulnerable Road User (VRU) safety is Spoke Safety. Spoke, in partnership with Qualcomm and automobile manufacturers such as Audi, have developed a solution using hardware and software to allow cars to “see” cyclists in near real-time, which can translate to significant reductions in VRU (think bicycles) fatalities. But before deep diving into CV-2X and how Spoke’s solutions may advance autonomous driving, it is important to understand what autonomous driving actually is and what is required to get there.

Self-driving cars are vehicles that can operate without human intervention using sensors, cameras, artificial intelligence and software. They are also known as autonomous vehicles or driverless cars. There are different levels of automation for self-driving cars, ranging from level 0 (no automation) to level 5 (full automation). Most of the current self-driving cars are at level 2 (partial automation) or level 3 (conditional automation), which means they still require human supervision and intervention.

Some of the technologies that are needed to make self-driving cars a reality include:

• Lidar: A laser-based sensor that creates a 3D map of the environment and detects obstacles and other vehicles
• Radar: A sensor that uses radio waves to measure distance, speed and direction of objects
• Cameras: Sensors that capture images and videos of the road, traffic signs, pedestrians and other vehicles
• Artificial intelligence: A system that processes data from sensors and cameras, recognizes patterns, learns from experience and makes decisions; and
• Software: A program that controls the vehicle's steering, braking, acceleration and navigation

Remember, Level 5 autonomous driving means that the vehicle can drive itself in any situation and environment, without any human intervention or supervision. This is the ultimate goal of self-driving technology, but it is also the most challenging to achieve for a number of reasons. 

Some of the critical technological gaps that prevent level 5 autonomous driving are:

• Software: Developing software that can handle complex and dynamic scenarios, such as bad weather, road construction, traffic accidents, pedestrians, cyclists and animals. The software also needs to be able to learn from its own mistakes and improve over time
• Sensors: Developing sensors that can provide accurate and reliable data about the surroundings, such as lidar, radar, cameras and GPS. The sensors also need to be able to fuse data from different sources and filter out noise and errors; and
• Infrastructure: Developing infrastructure that can support vehicle-to-everything (V2X) communications, such as roads, signs, signals and networks. V2X communications can enable vehicles to share information with each other and with the environment, enhancing safety and efficiency

What other companies besides Spoke are worth watching in this space and are at the cutting edge of autonomous driving? Well, there is no definitive answer to which company is the front runner in the development of self-driving cars as there is no universal criteria or metric to rank them (short of achieving Level 5 first). Some companies that are often mentioned, or are synonymous with autonomous driving include: 

• Waymo: A subsidiary of Alphabet GOOGL (Google's parent company) that has been developing self-driving technology since 2009. Waymo claims to have the world's most experienced driver with over 20 million miles driven on public roads and 10 billion miles driven in simulation
• Cruise: A subsidiary of General Motors GM that has been testing self-driving cars in San Francisco since 2013. Cruise aims to create a fully autonomous ride-hailing service that is safer, cheaper and more accessible than human-driven cars
• Tesla: which produces electric vehicles with advanced driver assistance systems called Autopilot and Full Self-Driving Capability. Tesla's TSLA vision is to achieve full autonomy without relying on lidar or high-definition maps, but rather on cameras and neural networks
• Daimler: owner of Mercedes-Benz MBGAF and other brands. Daimler has been developing self-driving technology under its Intelligent Drive program, which includes features such as Active Lane Change Assist, Active Brake Assist and Traffic Jam Pilot
• Ford F: who has partnered with Argo AI, a self-driving technology platform company, to develop autonomous vehicles for various applications such as ride-hailing, delivery and mobility services. 

But back to Spoke. Why and how can a company that is focussed on bicycle safety advance autonomous driving when there are other much larger companies investing in the space? Well, Spoke’s unique hardware and software allows municipalities to create a CV2X infrastructure at much lower cost than current solutions. Remember, infrastructure has to be present for the rapid scaling and adoption of new technologies. Without the investment into 3,4, and 5G infrastructure, smartphone adoption could not scale to what it is today. Without the investment into charging networks, electric vehicles cannot truly replace internal combustion engines as the electric car would be tethered to the home charging unit. In both these examples (and there are many), the up-leveling of infrastructure capabilities advanced the commercialization of the desired technology. In the case of autonomous vehicles, Spoke can help to up-level infrastructure and save costs by 10x, according to CEO Jarrett Wendt. Additionally, the ability to “see” VRUs in near real time is arguably superior to lidar, radar, cameras, and GPS because the devices seeking and speaking to each other is near-real time, which is essential in both being a front line in detection, but also a redundancy to the other sensing technologies mentioned. 

So how far are we from fully autonomous driving vehicles? The magic number from a variety of sources seems to be 2030, with infrastructure and regulation being some of the most important hurdles besides vehicle technology. Much like augmented reality and the Metaverse, there is significant momentum and inevitability in technologies like autonomous driving being a reality in our lifetime, and companies that advance the timeline with critical components such as sensing and AI are highly lucrative investment opportunities.