An Underground Garage Parking Space Layout Scheme Based On Exploration Strategy And Region Segmentation Algorithm

With the rapid development of our country’s economy,the number of urban vehicles is increasing continuously,the phenomenon of "driving difficult,parking difficult" has become very common.In order not to occupy urban land area,to alleviate the problem of urban traffic jams,the construction of a large number of underground parking has become an inevitable trend.Because the artificial design of underground garage will greatly test the technology of designers,but also requires a lot of time and energy,so many experts and scholars have carried out a lot of discussion on the problem of underground garage parking arrangement,a variety of automatic parking arrangement algorithms are constantly emerging,but most of these algorithms are only suitable for the parking lot with simple contour design and no obstacles inside.For underground garages with complex outer ring shape and irregular arrangement of obstacles,the existing automatic arrangement technology is not mature.Therefore,in order to arrange more parking Spaces under these complex conditions,this paper designs and implements a parking arrangement scheme based on exploration strategy and regional segmentation algorithm.Specific contributions are as follows:(1)Explore strategies to realize the laying of the main road of the underground garage,and preliminarily solve the connectivity problem of the drive way.This strategy mainly draws on the idea of reinforcement learning.In order to conveniently describe and record the state change of each position in the garage,the underground garage is first grid-cut according to the segmentation accuracy to generate the grid environment of the underground garage.The intelligent agent and the garage grid environment will constantly interact and pave the road.In this process,by setting a reasonable reward mechanism,the intelligent agent can choose the direction with the most rewards in the environment to explore,so as to maximize the accumulated rewards.(2)For the garage environment where the main road has been laid,the region segmentation algorithm is used to divide the open space.Through this algorithm,boundary area set and non-boundary area set can be obtained,and then different parking arrangement strategies are adopted in the two areas to ensure the optimal effect of parking arrangement in underground parking lot under the condition of lane connectivity.(3)In order to verify the effectiveness of the proposed algorithm,we use different standard CAD engineering drawings for experiments.The experiment mainly compares the proposed algorithm with the existing algorithm,and the results show that the proposed algorithm can effectively solve the parking arrangement problem under complex conditions.Through further analysis of the experimental results,it is found that the time consumption of the proposed algorithm is reduced by 40%on average and the number of parking Spaces is increased by 2.1%on average.At the same time,the number of invalid parking Spaces due to road congestion problems decreased.The results further show that the proposed algorithm can effectively help designers reduce workload and improve project income.