Transforma logo
Sign Up Login

Fully Autonomous Vehicles: Are they close to reality?

SEP 16, 2022 | Suruchi Dhingra
 
region: ALL sector: ALL Internet of ThingsHyperconnectivityHuman Machine InterfaceArtificial IntelligenceAutonomous Robotic SystemsEdge Computing

Earlier in September we published our report on the current status and future prospects for Autonomous Road Passenger Vehicles. In this blog post I share a summary of the findings of our research on what the major auto OEMs and other significant players in the space are up to, and how far we are from seeing widespread deployment of truly autonomous vehicles.

What are the major players up to in autonomous driving?

Many automotive manufacturers and automotive technology companies are making extraordinary efforts to develop a Level 3 or Level 4 autonomous vehicle but most of them are not quite there yet. Honda and Mercedes Benz are ahead of many; Mercedes Benz has legally been allowed to produce and sell a Level 3 autonomous vehicle (only in Germany) starting May 2022. The OEM will offer Drive Pilot (Level 3 autonomous driving system) in S-Class and EQS vehicles for EUR5,000 and EUR7,430 respectively. Elsewhere, Honda sold Legend vehicles equipped with Level 3 autonomous technology in 2021, but it was only a small batch of vehicles.

It's also worth noting that Tesla has made a lot of noise about autonomous vehicles and has made claims about being closer to launch of fully autonomous capability, but its vehicles still require constant driver supervision and are currently limited to Level 2 autonomy. Tesla has also not received regulatory approval yet for operating Level 3+ vehicles on road.

As a result of the lack of implementation of Level 3 or 4 autonomy by the auto makers, there aren’t many self-driving cars driven by consumers as personally owned vehicles on public roads yet. However, technology companies such as Waymo, Apollo Go, Auto X, Argo AI, Baidu, and Cruise have made good progress in the autonomous ride-sharing cars (with Level 4 capabilities) market in the US and China . Autonomous ride sharing cars or robo-taxi car services are widely spread across the streets of San Francisco, Arizona in the US, and around 10 cities across China including Beijing, Chongqing, Shenzhen, Wuhan, Shanghai, Guangzhou, Changsha, Cangzhou, Wuzhen, and Yangquan. In China, Baidu provided around 213,000 rides in the fourth quarter of 2021. The company plans to expand its services to 65 cities by 2025 and 100 cities by 2030. In contrast to this progress, there are companies that have received setback for concerns over safety procedures. Recently, Pony.ai lost its permit to test autonomous vehicles after finding violations on the driving records of some safety drivers.

Can the autonomous vehicle replace private car ownership? Ridesharing? Taxis?

This brings attention to the future of transportation in cities. The vehicle ownership status in the future remains unclear. Many studies assume that in future vehicles will no longer be privately owned and will be used with pooling options but there is currently little evidence to indicate whether this assumption is supported by user preference . Also, as cars are considered a status symbol for many people and there is also a general preference to own things rather than share which suggests that individual ownership will also exist with a significant share.

Autonomous Vehicles certainly have strong potential to dominate and expand the ride sharing market. One of the significant costs of conventional taxis, the driver’s wages, is removed. But the success of this opportunity is highly dependent on fleet volume restrictions by governments, any disruptions related to safety, and the potential for industrial action from taxi drivers.

What are the factors responsible for slow progress of autonomous vehicles?

Many OEMs had made claims of offering fully autonomous vehicles between 2020 and 2022 but weren’t successful as the road has been slippery and strewn with obstacles. A major roadblock that is delaying the launch of autonomous vehicles is lack of comprehensive regulations and concerns over safety. The current status of regulations is inconvenient for wider and efficient deployment of AVs. In the current scenario, there is lack of consensus on how to regulate the technology, which is mainly due to different views on liability ownership (in case of accidents), lack of differentiation between advanced driver assistance systems and AVs, and a potential of loss of jobs due to automation, amongst other factors. It is important to have a fair liability framework in place before launch or it can restrain or hold manufacturers from bringing partial automation or Level 3 capability in vehicles. For example, Audi cancelled its plans to launch Traffic Jam Pilot autonomous technology in A8 vehicles because of liability concerns as global regulators in Geneva hadn’t finalised the approval process for Level 3 vehicles at the time of its release.

Another hurdle that the market has to get through, is to overcome the high cost of hardware. The current cost of Lidar sensors is not at a suitable price point for mass deployment in passenger vehicles and definitely not a viable option for cheaper cars. Some companies have been able to bring down the cost of Lidar sensors with in-house production, but it is still very high. Companies can’t even compromise over the quality of Lidar sensors as it is the major component responsible for ensuring safe autonomy. A few companies have been trying to replace Lidar sensors with AI enabled cameras and radar sensors, but the concept is still at very early stages and doesn’t have clear evidence of success in near term.

Some other challenges that the market has to get through are the cost of autonomous vehicles and efficient process of transferring car control from computer to the driver in the event of emergency. Some manufacturers (such as Ford) are looking to skip Level 3 autonomy and go to Levels 4 or 5 directly, as they see the process of transferring car control from software to human driver, which is an element of Level 3, as a safety threat.

The result: a market slow to materialise

For all the reasons quoted above, and several others, the significant interest in autonomous driving will be very slow to translate into real world mass market deployments. According to our forecasts, we expect 46.6 million RGUs by the end of 2030, which will be only 2% of total road passenger vehicles.

2022

2021

2020

2019

All Blog Posts