Advanced Sensor Fusion for Visual Inertial Odometry in Dynamic Indoor Environments

工业机器人已经在自动化制造时代树立了重要的里程碑，在半个多世纪的时间里，工业机器人代替人类执行危险和可重复的任务。在这个世纪，专业和移动服务机器人将在工业和日常生活中扮演人类的同伴和照顾者的关键角色。与工业机器人类似，移动的服务机器人将承担服务任务中危险和枯燥的方面，释放人类服务工人的时间和认知需求，允许工人在高层决策和规划中做出更好的选择。在室内移动的服务机器人的上下文中，作为执行特定任务的先决条件，机器人必须知道如何定位自身和目标对象。因此，机器人必须配备能够处理社会环境中不确定性的导航系统。虽然存在基础技术，但很少发现足以满足现代需求。因此，本研究项目的范围是产生下一代视觉惯性测距机器人导航系统在动态室内环境。

Industrial robots have set important milestones in the era of automated manufacturing by carrying outdangerous and repeatable tasks in the place of humans for over half a century. In this century, professional andmobile service robots are going to play the key role as human counterparts and caregivers in both industrial anddaily-life setting. Similar to industrial robots, mobile service robots will take on the dangerous and dull aspectsof service tasks, freeing time and cognitive demand on human service worker, allowing the worker to makebetter choices in high level decision making and planning. In the context of indoor mobile service robots, as aprerequisite for performing a specific task, the robot must know how to locate itself and a target object. As aresult, the robot must be equipped with a navigation system that is capable of dealing with uncertainties insocial environments. Though foundation techniques exist, very few are found to be adequate in meetingmodern demands. Hence, the scope of this research project is to produce the next generation visual inertialodometry for robot navigation systems in dynamic indoor environments.