Waymo Deploys 29 Cameras, 5 Lidars for 'Superhuman' Safety as Tesla Backs Camera-Only AI
Waymo’s autonomy system uses 29 cameras, five lidars and six radars, targeting performance safer than human drivers while aiming to cut sensor count and costs. Tesla’s VP of AI asserts that camera-only vision captures enough data, framing autonomy as an AI challenge instead of a sensor issue.
1. Debate Context
Autonomous vehicle developers are divided over whether a multi-sensor suite or camera-only vision offers the best path to safe self-driving. This split underscores broader industry questions about the proper balance between hardware redundancy and AI-driven perception.
2. Waymo's Sensor Strategy
Waymo’s VP of onboard software highlighted a comprehensive sensor configuration—29 cameras, five lidars and six radars—to exceed human driving safety. The company plans to iteratively reduce sensor count and costs while maintaining or improving overall system reliability.
3. Tesla's AI-First Approach
Tesla’s VP of AI maintains that cameras already provide sufficient visual input, positioning autonomy as primarily an AI problem rather than a sensor limitation. The emphasis is on software refinement to extract critical information from standard camera feeds.
4. Uber Executive Perspective
Uber’s CEO Dara Khosrowshahi has stressed the challenge of achieving “superhuman levels of safety” with camera-only systems. He affirmed that rigorous safety standards must guide any path forward for robotaxis, regardless of sensor configuration.