Semantic map-based target localization and intelligent navigation
To address the issue of insufficient utilization of semantic information in complex indoor environments by mobile robots,a voice-controlled semantic navigation method based on a 2D semantic grid map was proposed.This approach fills the gap in the field of semantic navigation for indoor robots in single-scene environments.Utilizing the constructed 2D semantic grid map,a voice-controlled semantic navigation algorithm was employed to guide the robot to navigate near the corresponding semantic object based on voice commands.The robot accurately locates the semantic object and calculates suitable navigation points within a safe area around these objects.Combined with voice recognition and natural language processing technologies,the robot was capable of precisely moving to the specified semantic object based on vague user voice commands,achieving more intelligent and flexible navigation decisions.Experimental results demonstrated that the navigation success rate reached 93%,and in environments with irregular obstacles,the collision rate of this semantic navigation method was only 4.6%.The proposed voice-controlled semantic navigation method is feasible and exhibits significant advantages in obstacle avoidance.
NETL
NSTL
万方数据
