Dynamic Environment Navigation and Pathplanning for Assistive Robots using Nonlinear Predictive Control

本研究计画将针对一具医疗照护动机之移动的辅助机器人，提供特定路径规划与避障之解决方案。 这项研究是与Crosswing公司合作进行的，该公司开发了一种名为VirtualME的原型辅助机器人。 该机器人旨在通过将自主监控平台与复杂的通信系统相结合来为老年人提供服务。 VirtualME项目的目标是帮助老年人在自己的环境中更独立地生活，并进行安全监测，同时减轻人口老龄化对医疗保健系统的压力。 它面向放置在私人家中，并意味着与其环境和远程监督者进行自主互动。 其成功的关键需要是它能够在具有已知结构但未知移动障碍物的环境中移动。 此外，至关重要的是，在运动时，它可以与它所服务的老年人或其环境中的指定对象保持最大的视觉接触。 此外，当潜在的移动障碍物对给定的路径产生阻碍时，机器人必须能够自主地重新定向其路径以完成其任务。目前，VirtualME只能以固定方向在预先编程的方向上运动，并且不能与其环境自主交互。一些预测控制路径规划的基础已被纳入，但与移动障碍物和路径规划中的相互作用是不可用的。 本研究将发展一非线性预测控制器来完成路径规划与避障。 这将使VirtualME能够以安全的方式放置在有移动障碍物的环境中。将这些功能扩展到当前的VirtualME平台将是一项重大成就，这将满足持续开发Crosswing的移动的辅助技术的关键需求。 在所描述的环境中实现自主控制之前，无法完成用于继续开发的VirtualME机器人的部署。

This research project will provide solutions for specialized path planning and obstacle avoidance for a healthcare motivated mobile assistive robot. The proposed research is in collaboration with Crosswing Inc, who have developed a prototype assistive robot called VirtualME. This robot is designed to serve the elderly population by combining an autonomous monitoring platform with a sophisticated communication system. The goal of the VirtualME project is to help the elderly live more independently in their own surroundings with safety monitoring while simultaneously alleviating the strain of an aging population on the healthcare system. It is geared towards placement in a private home and is meant to interact autonomously with its environment and with a remote supervisor. A critical need for its success is the ability for it to move around an environment with a known structure, but unknown moving obstacles. Furthermore, it is critical that while in motion it can maintain maximal visual contact with the elderly person it is serving or with a specified object in it's environment. In addition, when potentially moving obstacles create a block to a given path, the robot must be capable of autonomously redirecting it's path to accomplish it's tasks. Currently the VirtualME is only capable of motion in a preprogrammed direction with fixed orientation and does not interact autonomously with its environment. The foundation for some predictive control path planning has been incorporated, but interaction with moving obstacles and in the path planning is not available. This research project will develop a nonlinear predictive controller to accomplish the path-planning and obstacle avoidance. This will give the VirtualME the ability to be placed into an environment with moving obstacles in a safe manner. Incorporating these features into the current VirtualME platform would be a major accomplishment, which will address a critical need for the continued development of Crosswing's mobile assistive technology. Deployment of the VirtualME robot for continued development cannot be accomplished until autonomous control in the described environments has been implemented.