Dynamic Path Planning Method And Simulation Research Based On Video Data

In past decades,pedestrian evacuation has become a popular issue in society.Since the reform and opening up,China has experienced unprecedented economic explosions,populations have continued to integrate into cities,urban density has continued to grow,and more and more supermarkets have entered the daily life of people.These supermarket and cultural squares provided people with great convenience while brought new threat to safety of inhabitant.The number of tourists in these places tends to be large and concentrated.In the event of an emergency such as a fire,without reasonable guidance,visitors will suddenly become confused and panic-stricken to cause stampede events,thus threatening their lives and property.In order to reduce or even avoid these crises,scholars have been conducting research on crowd evacuation simulation.Crowd evacuation simulation refers to the use of computer modeling to simulate the evacuation behavior of people in real situations,so as to know the possible evacuation of pedestrians in advance and avoid risks.Studies have shown that the use of computers for crowd evacuation simulation has many outstanding advantages and has more constructive significance.In the crowd evacuation process in complex scenarios,macro path planning navigation is a very important factor.At present,most of the researches either lack macroscopic path planning,or the path planning method is not real enough and the calculation amount is huge.In view of the above shortcomings,this paper realizes the fusion of video data and improved social force model by constructing a strategic layer-behavior layer hybrid model.The pedestrian flow trajectory is extracted from the real evacuation video,and then discretely optimized and stored in the hierarchical database to provide a basis for the path planning as a strategic layer model.In the traditional social force model,the gravitational and follow-up effects of the group are added,and the isolated individuals,groups,and pedestrian flow phenomena composed of isolated individuals and groups are considered.The improved social force model is used as the behavioral layer model to realize the movement in complex scenes.We propose a method of using differential evolution algorithms to optimize the shape of obstacles,seeking the optimal obstacle shape to maximize the evacuation efficiency.Aiming at the congestion problem at the vertices of obstacles in complex scenes,the concept of density domain bypass points is proposed based on global path planning.Local path planning leads the pedestrian flow to bypass the high-density area of the crowd and realizes the optimal path in real time.The main work of this paper is as follows:1.Since the traditional evacuation model motion path is completely calculated by the function,its evacuation effect is inevitably lost.Moreover,in order to reduce the complexity,based on the data-driven technology,this paper proposes a dynamic path planning method of the strategic layer-behavior layer double-layer hybrid model to realize the fusion of video data and improved social force model.2.The traditional social force model,pedestrians move to the destination in isolation,without communication between each other,obviously,this is different from the real world.Taking full account of the diversity of pedestrian composition in real life,We define three different kind of pedestrian evacuation: isolated individuals,groups,and pedestrian flows consist of isolated individuals and groups.Then we improved original social force model by adding the gravitation and follow-up effects within the group.3.As for the obstacle problem in path planning,The suitable location and geometry of obstacles can effectively improve the evacuation efficiency,Thus we proposed a method which utilize differential evolution technique to optimize the shape of obstacles,the best obstacle shape is determined by this method.At the same time,in order to overcome the problem of congestion when pedestrians walk around obstacle vertices,a dynamic path planning method of density-domain detour points is proposed based on global path planning.Local path planning guides pedestrians to bypass high-density areas of the crowd and chooses the optimal path dynamically in real time to solve the congestion problem at obstacle vertices in complex scenes.Finally,the above-proposed method is applied to the crowd evacuation simulation platform to realize the pedestrian flow evacuation simulation system,which can perform scene modeling,path planning,and real-time evacuation of pedestrian flow.We conducted simulation studies on the teaching hall and campus scenes.Summarize the impact of factors such as the number of pedestrians,the number of exports,congestion,and distance on evacuation.By comparing the original model with the model proposed in this paper,the proposed model has better simulation degree,realizes rapid pedestrian evacuation,and the evacuation effect is more in line with the actual situation,which plays an important role in guiding the evacuation of people in a sudden situation.