Autonomous Vehicles

Autonomous vehicles (AVs), also known as self-driving or driverless vehicles, are road vehicles that use artificial intelligence, sensors, mapping and onboard computing to operate with limited or no human involvement. The field spans passenger cars, robotaxis, delivery vans, buses and logistics equipment, and has moved from research laboratories into commercial services in a growing number of cities, led by operators such as Waymo, Tesla and China's Baidu Apollo Go.^[1][6]

History and Background

Automated driving research began in the 1980s, and the field accelerated in the 2000s through the United States Defense Advanced Research Projects Agency (DARPA) Grand Challenges, competitions that demonstrated self-driving vehicles navigating desert and urban courses. Google started its self-driving car project in 2009, which was reorganised as Waymo in 2016 and later became the first operator to run fully driverless paid services at scale, beginning with Phoenix, Arizona and expanding across the United States. Tesla pursued a different path, shipping progressively more capable driver-assistance software to consumer cars and eventually an unsupervised robotaxi service using the same vision-based technology stack.^[1][6]

After a decade that mixed technical breakthroughs with disappointment, the mid-2020s brought commercial momentum. Waymo expanded paid, fully driverless service to 14 metropolitan areas across the United States by September 2026, including Denver, San Diego and Tampa, with a fleet of roughly 4,000 vehicles; Tesla operated paid robotaxi services in several Texas and Florida metros and introduced the steering-wheel-free Cybercab in Austin; and Amazon's Zoox began charging for rides in Las Vegas. Chinese operators scaled in parallel, with Baidu's Apollo Go running driverless services in around 20 cities across China and reporting millions of fully driverless rides in the first quarter of 2026.^[1][6][7][8]

Key Concepts and Technology

Regulators and engineers classify automation using the SAE International standard J3016, which defines six levels: Level 0 for no automation; Level 1 and 2 for driver assistance and partial automation, where a human still supervises; Level 3 for conditional automation under defined conditions; Level 4 for fully driverless operation within a defined area or operating domain; and Level 5 for automation that matches or exceeds human capabilities everywhere. Most commercial robotaxi services today operate at Level 4, meaning they can drive without a human aboard but only inside mapped areas and subject to remote monitoring.^[9]

Two broad technical approaches dominate. Sensor-rich systems, such as Waymo's, combine lidar, radar and cameras with detailed high-definition maps and heavy validation, allowing operation in complex cities at the cost of mapping each new market. Vision-centric systems, such as Tesla's, rely primarily on cameras and neural networks trained end-to-end on large datasets of driving, with generalisation traded against supervision requirements. Both rely on machine learning for perception, prediction and planning; remote operations centres, which can assist a driverless vehicle when it encounters an unusual situation, have become a standard part of commercial deployments and a key focus for regulators evaluating safety.^[1][2]

Applications and Impact

The most visible application is the robotaxi — ride-hailing in a driverless vehicle. Independent trackers counted 14 United States metros with public Waymo service as of September 2026, alongside Tesla's robotaxi markets and newer entrants such as Zoox. The Stanford AI Index reported that driverless rides grew about 175 per cent year over year in 2025-2026, one of the clearest signals that autonomous mobility has entered commercial scale. Beyond passenger transport, AVs are being applied to freight and last-mile delivery, port and warehouse logistics, and closed-campus shuttles.^[1][5][6]

The technology's impact is debated. Proponents point to the potential to reduce road deaths caused by human error, lower the cost of transport and logistics, and free drivers from repetitive routes; sceptics raise incidents involving early systems, the employment implications for driving occupations, and questions of legal liability when a vehicle with no driver causes harm. Regulation has consequently become a competitive factor, with jurisdictions that create clear permitting pathways — such as Nevada's 2026 approvals for multiple robotaxi operators — attracting investment and trials.^[6]

>See Also

• Physical AI
• Computer Vision
• Cyberjaya
• AI in Malaysian Logistics
• Edge AI
• Reinforcement Learning