M14 Intelligence’s recent global analysis on autonomous vehicles industry predicts that the demand for level 2 passenger vehicles will grow exponentially to more than 21 million units by 2030 and cross 40 million unit mark by 2040.
At present, 4 in 10 consumer purchase cars are equipped with some kind of advanced safety feature including automated emergency braking, adaptive cruise control, lane assist systems, and forward collision avoidance among others. Majority of OEMs are offering these safety and warning systems as a standard feature in their cars. However, automation in the form of hands-off driving, automatic braking, and acceleration or automatic lane change or even automatic valet parking features are still offered on higher trim levels of premium models at an additional cost to consumers. In total, less than 2.5 million cars today are equipped with level 2 automation features by brands such as Tesla, Audi, Mercedes Benz, BMW, Toyota, Nissan, Volvo, Ford, GM, FCA, PSA, and Hyundai among others.
The competition among the OEMs is intensifying.
From the operational point of view, the operational design domains of each OEMs even within their trim levels vary a lot. The highest trim offers advanced systems that can control brakes, steering, and acceleration on its own, which assist the driver in changing lanes, stay at safe distance from the vehicles around, and accelerate at free ways. Such features are majorly marketed in the form of level 2+ or level 2++, which is ideally a more of marketing gimmick by the OEMs. With two clear reasons - one to justify the cost of the vehicle and two to showcase edge over its competition. Depending on the highest autonomy feature offered on a trim-level, the research categorizes vehicles under SAE’s defined levels of autonomy - ideally level 1 or level 2, looking at the current autonomy status.
The next obvious goal for OEMs is to achieve level 3 autonomy in the form of highway autopilot and remote automated parking. Even though the systems are being developed and some OEMs are ready with such solutions, the biggest challenge the industry is currently facing is to market such features.
Biggest challengeis in the form of regulations, especially in United States and Europe. The regulatory frameworks are still not in place in both these regions for OEMs to market their vehicles with level 3 autonomy. The best example is Audi’s A8 with Traffic Jam Assist, which the company has marketed as world’s first level 3 car. However, due to lack of regulations in place for this level of autonomy, Audi waited for 3 years since mid-2017 and finally called off their plan for level 3 autonomy. OEMs are now cautious and intentionally not marketing their vehicles as a level 3 car rather renaming it as level 2+ or level 2++.
In the east the situation is changing rapidly. China being one of the most promising markets for advanced featured autonomy solutions. Chinese government and regulatory bodies are pushing the move to commercialize level 3 and above autonomy along with electrification of vehicles. This will certainly boost the prospects for the local companies right from the OEMs to technology developers. According to M14 Intelligence estimates, more than 7.3 million cars equipped with level 2 automation will be shipped by 2030 in China alone, becoming the single largest market globally. While the cars equipped with level 3 systems will witness sales of around 1.5 million units by 2030.
The true self-driving comes with the complete hand-over of the vehicle from driver to the system. Even though all the OEMs are extensively working to achieve full autonomy only few have made significant development in this space. Besides, the market for full autonomy is not in the consumer purchase vehicles but in the shared mobility solutions. There are two major reasons for this -
An expensive solution
For enabling a vehicle to drive autonomously the system needs to be robust. The system needs multiple sensors for perception of environment (including camera computer vision, radars and LiDAR sensors too), precise mapping and localization using navigation and HD maps, precise decision making machine learning algorithms or artificial intelligence, and connected infrastructure for vehicles to communicate with each other and the infrastructure. The hardware and software requirement to build such systems costs heavily which will in turn have to bear by the consumers if integrated in a private purchase vehicles. So the best possible way to monetize, is to enter on large scale deployment of autonomous vehicles for shared mobility services like robotaxis, autonomous shuttles, autonomous pods, or even in self-driving trucks for long-haul platooning. This is clearly visible by the strategies adopted by the leading OEMs in the automotive market. Almost every OEM has either partnered or acquired a shared mobility fleet company or a self-driving solutions company to develop and deploy highly autonomous driving. Volvo, FCA, Toyota, JLR partnered with Waymo; GM acquired Cruise; Ford and VW has stakes in Argo AI; Lyft has partnerships with Aptiv, BMW, Toyota, and Volvo; Hyundai built its self-driving company called Motional in partnership with Aptiv and also has previously partnered with Aurora; Uber has partnerships with Volvo and Toyota, Baidu’s Apollo has partnered with more than 50 Chinese, American, and European automakers. The list is endless, the industry is getting together forming consortiums to together to achieve the common goal of self-driving vehicles.
Stakes are high
It is important to understand that a vehicle has to drive autonomously in a mixed environment of todays roads where it will be a mix of traditional vehicles (with no safety features to some or advanced safety features), pedestrians, cyclists, and many other factors. In such environments, system has to face numerous scenarios that might be untested or not virtually simulated before or witnessed by the AI system before. The system must perform 100 percent and there is no room for error, as in case of any accident, the liability lies with whom? On contrary, robotic vehicles such as robotaxis, shuttles, and pods have operational domains designed to drive such vehicles at a limited speed (also called as slow moving urban autonomous mobility). Initial deployment of such vehicles will be limited to geo-fenced areas. Best examples for such deployment is Waymo One (Pheonix) and Baidu’s Apollo Go (Haidian and Yizhuang, Beijing) robotaxis. Both these companies has initiated their self-driving robotaxis fleet without safety driver. While Cruise has recently received permit from DMV to test their autonomous vehicles without the presence of a safety driver. It is important to note that all these services are provided on a mapped and geo-fenced areas only, because the systems are not currently trained to explore new roads on its own.
Tesla is getting aggressive!
Tesla electric cars grabbed top ratings in the US, Europe and Australia from vehicle and highway safety authorities, however the most recent testing in Europe, shows Tesla’s Autopilot needs to work more on its driver engagement. This echoed in a recent court ruling in Germany too, that resulted in banning Tesla Germany from advertising a “full self-driving capability” or “autopilot inclusive” in its marketing materials, as according to the court its misleading. Having said that, Tesla is one the biggest contenders in the race to full autonomy. This is crucial considering the Tesla’s systems do not incorporate LiDAR sensors. Tesla believes that LiDAR is unnecessary and expensive for autonomous driving while computer vision system along with radar and ultrasonic sensors are enough to self-drive. Tesla’s self-driving system can drive on unexplored roads as it is completely based on its machine learning computer vision system. Tesla claims to have registered one accident for every 4.53 million miles driven in which drivers had Autopilot engaged in Q2-2020. However, Tesla is clear on its choices to attain full self-driving and claims to launch a level 5 self-driving system that can be used for robotaxi service by end of 2020.
In our interview in June 2020 with former Tesla employee (confidential source), Tesla’s computer vision ML algorithms are much advanced and continuously evolving with the road scenario information being collected from existing Tesla vehicles on roads across the world. The source also claimed that Tesla’s systems can outperform many existing high-end premium vehicle autopilot offerings in many ways. Although these are strong claims, with self-driving the stakes are very high, and its very important that the system design has high safety measures in case of unexpected scenarios.
To read more please download free abstract of our recently published report on Autonomous Driving Industry. This recently published research analyses all the leading level 2 (including level 2+ and level 2++) systems, their operational design domains, and active safety features offered by 70 OEM brands across globe. The research also gives analysis on the self-driving systems, its development (including level 3 and level 4 - highway driving and robotaxis, shuttles, and pods) and deployment plans of 80+ OEM brands. The report also analyses the complete ecosystem to map who partners with whom and who supplies to whom.
M14 also published numerous studies with deep dive analysis on individual autonomous driving enablers including camera systems (in-cabin and world facing), radar, LiDAR, GNSS, HD-maps, annotation, and simulation among others. Please check INDUSTRY REPORTS page for more available off-shelf research reports.
autonomous vehicles market
level 2 cars
Autonomous Vehicles and Technologies Market
Read in-depth analysis on autonomous vehicle development maturity on the leading OEMs and self-driving solutions providers