Autonomous vehicles are moving from virtual tests to early development
Autonomous vehicles (AVs) are moving from virtual tests to early development phases worldwide in commercial spaces like ride-hail services, freight logistics, and mobile retail. Both consumer and commercial vehicles are leveraging real-time data to build complex simulations and understand how artificial intelligence (AI) responds to unpredictable situations, says GlobalData, a leading data and analytics company.
Manish Dixit, Practice Head at GlobalData, comments: “AVs are focused on using advanced algorithms and sensing technology to improve vision, radar, Light Detection and Ranging (LiDAR), and vehicle control systems. AI is helping in path planning and predictions to forecast the behaviour of other vehicles and pedestrians, and with predictive maintenance to reduce downtime and maintenance costs of AVs.”
Prathyusha Paruchuri, Senior Disruptive Tech Analyst at GlobalData, comments “AI is an integral part of AVs functionality, starting from observation and decision-making to control and supervision. AI can not only reduce traffic congestion and optimize travel routes but also eliminate human intervention in driving. However, it still needs to overcome challenges in terms of safety and reliability of AVs and improve overall vehicle performance.”
The outbreak of the COVID-19 pandemic along with a slowdown in economic development and global funding has brought many AV trials to a temporary halt. Moreover, the high investment costs, lack of profitability, and the time involved in manufacturing AVs is leading to the exit of many players. For instance, Uber disposed of its AV unit in late 2020, and Argo.ai joint venture between Ford and VW was dispersed in 2022. Subsequently, companies like Tesla, Waymo, and Baidu are looking to manufacture safer and more reliable vehicles.
GlobalData’s latest report, “Digital Innovation Case Studies in Automotive,” highlights the implementation of AI technologies with real-world use cases and assesses the return on investment.
Continental leveraged AI to reduce time-to-market
Continental AG, a Germany-based supplier of automotive parts, collaborated with NVIDIA, a provider of AI and IBM Spectrum storage for data. They launched Advanced Driver Assistance Systems (ADAS), which use deep learning and trained artificial neural networks to develop intelligent sensors and data-driven traffic safety solutions. The systems are claimed to improve AI training time by 70 per cent and reduced training time from weeks to days.
BMW uses big data and machine learning to boost productivity
BMW Group collaborated with DXC Technology, an IT and consulting services provider, to effectively access and analyse the vast amounts of data required to develop and improve the driving algorithms for its automated vehicles. The platform provided nearly 230 petabytes of usable storage and established sufficient processing capacity to simulate up to 240 million kilometres of test data within a period of just 3 months.
Honda harnesses AI to improve operational efficiency and profitability
Splunk, a technology provider, was approached by Honda Manufacturing of Alabama (HMA), which is Honda’s biggest light truck production facility, to implement its machine learning (ML) solution. Splunk claims that its ML solution allowed HMA to access the right data and reduce the mean time to repair (MTTR) by 70 per cent.
Paruchuri concludes: “The use of AI and machine learning is converting the vehicles into intelligent and adaptable systems. AVs are also incorporating cybersecurity to safeguard passenger data from malicious activity. Manufacturers are now focusing on Generative AI that uses synthetic training data to conduct virtual tests in a controlled environment, which could train their AVs more efficiently and effectively.”
