Experience: Implementation of Edge-Cloud for Autonomous Navigation Applications

Autonomous navigation, while still facing a lot of challenges, has become a reality in the last few years. It has been successfully deployed in limited environments, showing the potential such technologies offer. The availability of increased computational power, coupled with advances in machine learning techniques, including deep learning, have enabled this success. However, there are many challenges that need to be overcome to enable massive adoption of autonomous navigation in all environments. One such challenge is to provide reliable, low-latency, and cost-effective data processing solutions for compute-heavy applications. To address this challenge, data processing near the data source, that is, at an edge cloud has been proposed. In this paper, we share our experience in implementing one such edge cloud, designed to bring compute power close to resource-constrained end devices in an autonomous navigation testbed, named as Technology Innovation Hub on Autonomous Navigation and Data Acquisition Systems (TiHAN). By considering some use cases, we show how deploying this edge cloud at the testbed has greatly benefited the performance of autonomous navigation applications.