IEEE will make a standard for safety-considerations in automated vehicle (AV) decision-making, the first version of which it intends to publish within a year. The workgroup will be led by Intel Senior Fellow Jack Weast and will take Mobileye’s RSS model as the starting point, Intel declared on Thursday.
The Institute of Electrical and Electronics Engineers (IEEE) has approved the offer for a model in automated car decision-making. It intends to develop a rules-based, mathematical model that shall be technology-neutral and verifiable via math. It is also going to be adjustable to allow for regional customization by local governments.
The proposed standard may also embody a test methodology to evaluate conformance with the standard.
The necessity for such a standard has emerged from the need to manage and assess the safety of AVs. Some automotive SoC manufacturers are promoting hundreds of TOPS for their systems. Those chips run deep learning artificial intelligence algorithms.
While self-driving methods are compute hungry, a deep learning-based method creates an effect on a black field that can solely be examined by statistical proof via perhaps millions of miles of driving or simulation.
A formal guidelines-based model would offer an alternative means to confirm AV decision-making and to make sure the safety of self-driving programs versus those last resort statistical arguments.
Perhaps proof that that is necessary, GM Cruise and Daimler delayed their plans for robotaxis this year, based on security concerns, which proved to be a harder problem than the firms thought.