A method of autonomous navigation under space-time map based on multi-target tracking and prediction

Abstract

The invention discloses an autonomous navigation method under a space-time map based on multi-target tracking and prediction. The method obtains a two-dimensional spatial occupancy grid semantic map and a time series of target positions, and performs information on the two-dimensional spatial occupancy grid semantic map according to the time dimension. Stacking to obtain an xy-t map; project the target position time series onto the xy-t map; perform path planning in the xy-t map to obtain the optimal path; the planning process needs to follow the principle of time irreversibility while avoiding all dynamic targets And the influence of the static environment; the optimal path is divided into multiple segments according to the time step and mapped to the xy plane, and the two-dimensional trajectory sequence and continuous expected velocity sequence in multiple space domains are obtained. Autonomous navigation is accomplished according to the two-dimensional trajectory sequence and the continuous desired velocity sequence in the space domain. The invention not only effectively overcomes the disturbance of the dynamic target to the path planning, but also, compared with the traditional space navigation map, the space-time navigation map can greatly simplify the dynamic planning problem.

Description

technical field [0001] The invention belongs to the technical field of robots, and in particular relates to an autonomous navigation method under a space-time map based on multi-target tracking and prediction. Background technique [0002] Intelligent Vehicle (Intelligent Vehicle, IV), also known as Unmanned Ground Vehicle (UGV) or Wheeled Mobile Robot (WMR), is a set of functions such as environmental perception, planning and decision-making, and multi-level assisted driving. The comprehensive system in one has extremely high socio-economic value and national defense military value. [0003] At present, the research and application of intelligent vehicles mainly focus on the urban structured environment that relies on prior information such as geographic information systems or high-precision maps, and there are relatively few researches on autonomous navigation technology in highly dynamic unknown areas. The latter is of great significance for the application of intelligen...

AI Technical Summary

Problems solved by technology

[0005] At present, the path planning of the existing technology is implemented in two-dimensional or three-dimensional maps, only considering the static environment, and cannot overcome the impact of random movement of various dynamic targets

If the consideration of dynamic goals is added to the planning process, dynamic programming will increase the difficulty of planning

Images