Communication Challenges and Solutions for Distributed Battery Management and Second Life Arrays

Abstract

In the past, distribution of signals within traction batteries has been carried out with wires. For example, lengthy voltage cell sense leads for lithium batteries ran from cell terminals to centralized controllers. The industry is moving to wireless data to replace wiring. A key enabler to wireless powertrain networking is the ability to exploit redundancies within multiple signals so as to reduce the required networking bandwidth. We provide an approach for bandwidth reduction which uses unsupervised learning in order to develop the training sets for long short term memory that reduces the number of signals transmitted.