Energy efficient machine learning based SMART-A-BLE implemented Wireless Battery Management System for both Hybrid Electric Vehicles and Battery Electric Vehicles

Abstract

Electric vehicles have a larger battery pack with more than 100's cells connected in series or parallel combinations to form many battery modules. A Battery Management system (BMS) continuously monitors Current, Voltage, and Temperature data to trigger control algorithms like cell balancing and thermal management using wired communication which is then sent to the cloud for real-time analytics [1]. Wired communication restricts the flexibility in battery module assembly and makes it difficult for battery swapping. The proposed wireless communication, along with the SMART-A-BLE algorithm for BMS in the electric vehicle, uses Bluetooth Low Energy [2], [3], [4], and [5]. SMART-A-BLE algorithm predicts the best connection interval using the Decision tree model based on the link quality factor captured based on real time latency. The model uses threshold latency, expected retransmission factor, data packet length, and the number of peripherals as input features. A look-up table-based algorithm is also proposed for resource constraint devices to achieve the optimal operating point where the slaves could transmit at both low latency and low current consumption with less congestion. SMART-A-BLE control algorithm outperforms the mathematical model-based algorithm in both ideal and interference conditions within a metallic container by maintaining robust connectivity.