Abstract: In the intelligent inspection robot for mine air compressor units,when traditional algorithms are used for the
path planning of intelligent inspection robots,they encounter problems such as slow optimization speed and easy trapping in local
optimal solutions in complex mine environments. To improve the inspection efficiency and accuracy,an Improved Bat Algorithm
(IBA) is proposed. Firstly,a uniform initialization strategy is adopted to ensure that the initial positions can widely cover
the decision space. Secondly,in the iterative update process,a golden sine operator is introduced to optimize and update the bat
individuals that perform well in the fitness evaluation,while the average position of the population is used to guide some individuals
to narrow the search range while maintaining a fast convergence speed. Additionally,a dynamic inertia weight coefficient
is introduced in the global search stage,and a search strategy combining single-dimensional and full-dimensional is adopted.
Experiments show that under 5-dimensional conditions,the IBA algorithm converges in only 20 iterations in the Sphere function
test,far fewer than the Bat Algorithm (BA),and performs well under 50-dimensional conditions. In terms of the path planning
effect of the robot,the path planned by the IBA algorithm is shorter than that planned by BA,Adaptive Bat Algorithm (ABA),
and Global Chaos Bat Algorithm (GCBA),and has fewer turning points,fewer convergence iterations,and lower fitness values in multiple scenarios. The research reflects that the intelligent inspection robot path planning method based on the IBA algorithm
can increase the inspection efficiency of mine air compressor units by 45. 9% and improve the fault detection accuracy to
98. 9%. The proposed algorithm is helpful for the efficient path planning of intelligent inspection robots for mine air compressor
units and contributes to the safe production of mines.

Key words: mining air compressor unit,intelligent inspection,path planning,bat algorithm,golden sine