Distributed Optimization of Field Sensors Guided Autonomous Vehicular Operations

点击翻译按钮获取中文摘要

Autonomous robotic vehicles, ground or aerial, play critical roles in various intelligent operations seen or desired in our current and future lives, ranging from home services and industrial manufacturing, to more and more field applications such as natural resource monitoring, pipeline and building inspection, precision agriculture, mineral exploration, search and rescue operations, etc. Due to the increasing complexity in functioning expectation, there is no doubt that networked and collaborative operation of autonomous robotic vehicles will be in high demand. It is well known that one of the biggest challenges facing autonomous vehicles is the real-time sensing and perception of their surrounding environment. The current sensing technologies facilitate the broad application of various field sensors such as cameras, LIDAR/RADAR, ultrasonic sensors, etc. However, due to the nature of field perception and measurements performed by these sensing devices, which essentially is an infinite-dimension problem in space, huge challenges are posed for effective and efficient applications of them, especially, when applied in autonomous operations which normally demand tremendous robustness of performance. Another major concern in practices about field sensing devices is the cost of devices themselves, whereas the realities in autonomous operations call for solutions of networked low-cost sensing units instead of powerful single or multiple high-cost units. Apparently, it would be of great significance for both research and industrial applications to dig out the greatest potential of field sensing and perception as applied in various autonomous operations (for example, autonomous operation of ground or aerial robotic vehicles) so that these operations could be delivered in a way that is both the most effective and the most cost efficient. This research proposal intends to address the distributed optimization for field sensing networks with the operation of autonomous robotic vehicles being the showcase of application. It is noted that the focus will be on developing a systematic design framework which will be featured with novel modeling thinking, new concepts of problem formulation, practically feasible design procedures and algorithms, and meaningful applications to autonomous robotic operation tasks. In particular, it is expected that the sensing functions and robotic vehicle operations are integrated to generate conceptually new methods to address the said problems and to design robust and optimal control and management mechanisms for the networked autonomous robotic vehicles. It is also argued that the framework developed would be highly industrial-practices oriented and hence would find its way to be easily applied in practices.