Toyota's adoption of TIER IV's end-to-end autonomous driving technology highlights how an open strategy is accelerating participation by other companies and data sharing. TIER IV is advancing collaborative AI training through Co-MLOps under an open-source strategy centered on Autoware.
Toyota-adopted TIER IV end-to-end autonomous driving is based on an “E2E” model that goes beyond conventional modular development, with a single large-scale AI processing everything from the in-vehicle camera footage used as input to vehicle control as the output. This is expected to eliminate the loss associated with the relay-style transfer of data in conventional systems and enable driving with greater flexibility and adaptability.
At the core of its open strategy, TIER IV is promoting Autoware, described as the world's first open-source autonomous driving software. It has proposed a framework in which multiple companies can jointly train and improve autonomous driving systems using shared data through a collaborative machine-learning platform known as Co-MLOps. Major component manufacturers, including Hitachi Astemo, are also pursuing capital and business alliances, with mass production using the open platform in view.
Compared with the vertically integrated Tesla model, the open ecosystem is seen as capable of gaining a competitive advantage from the interaction between data volume and computing power by harnessing the “gravity” of data. At the same time, practical hurdles such as the black-box nature of the technology, authorization for use on public roads, and the allocation of liability are also identified. A hybrid design is proposed as the practical solution for widespread adoption.