March 10 (Reuters) – Qualcomm and British self-driving startup Wayve said on Tuesday they are collaborating on an integrated artificial intelligence system \u200cto help automakers rapidly deploy advanced driver-assistance and automated driving features.<\/p>\n
The \u200ctie-up combines Wayve’s “AI Driver” software with Qualcomm’s Snapdragon Ride automotive chips and active safety software, \u200bcreating a platform for carmakers to use across models ranging from entry-level systems to advanced automated driving capabilities.<\/p>\n
Chipmakers and software developers are racing to supply the technology for future vehicles, as automakers seek systems to accelerate the \u200crollout of increasingly automated driving \u2060features.<\/p>\n
The companies said the integrated system aims to reduce the complexity automakers face when stitching together chips, safety systems \u2060and AI software<\/a> from multiple suppliers, and will support features from hands-off assistance to advanced “eyes-off” driving functions as regulations allow.<\/p>\n The combined system is designed to scale \u200bacross \u200bvehicle tiers and geographic markets, allowing carmakers \u200bto standardize underlying technology, the \u200ccompanies added.<\/p>\n Nvidia-backed Wayve develops an AI model using real-world driving data, enabling vehicles to learn driving behavior and adapt to different road conditions and regions without extensive rule-based programming.<\/p>\n Qualcomm, which has been expanding beyond smartphones, said its Snapdragon Ride platform provides the high-performance, energy-efficient processing for advanced AI systems \u200cin vehicles while meeting safety standards.<\/p>\n Automakers have \u200bshown growing interest in systems that can \u200bshorten development cycles and allow \u200bsoftware updates to expand capabilities over a vehicle’s lifetime, \u200cthe companies said.<\/p>\n Wayve, founded in 2017, \u200bis part of a \u200bnew wave of AI-focused autonomous driving developers pursuing software-centric approaches that rely on machine learning rather than heavily map-dependent systems.<\/p>\n