Supporting Deep Learning Based Autonomous Driving Applications on Resource-Constrained Devices

With recent advances in computing and sensing technologies, autonomous driving has gained increasing interest and become a promising platform to support the next generation intelligent transportation systems. A critical requirement for autonomous driving systems is to be able to utilize AI and machine learning techniques to make reliable decisions on edge devices in a timely manner. Deploying reliable machine learning models on edge devices in a real-time environment is a challenging task. Real-time applications such as traffic surveillance or traffic sign detection require consistently low latency in order for the device to keep up with its environment. Edge devices are able to compute machine learning tasks without needing to offload computation to a cloud server, however they often have a tight limit on the amount of available resources. Quantization is a model optimization technique used to increase inference speed at the cost of accuracy. In this paper, we carried out a comprehensive study on utilizing quantization to enable efficient and effective decision making for autonomous driving applications. Various quantization methods are analyzed, with a focus on the trade-offs between the learning performance and computational performance. Furthermore, we identify pre-processing techniques and types of data input that may improve the model’s accuracy. The proposed approach will enhance the decision making process in resource-constrained environment, and provide more flexibility for deploying the autonomous driving applications.