Existing crowd evacuation guidance systems require the manual design of models and input parameters, incurring a significant workload and a potential for errors. This paper proposed an end-to-end intelligent evacuation guidance method based on deep reinforcement learning, and designed an interactive simulation environment based on the social force model. The agent could automatically learn a scene model and path planning strategy with only scene images as input, and directly output dynamic signage information. Aiming to solve the “dimension disaster” phenomenon of the deep Q network (DQN) algorithm in crowd evacuation, this paper proposed a combined action-space DQN (CA-DQN) algorithm that grouped Q network output layer nodes according to action dimensions, which significantly reduced the network complexity and improved system practicality in complex scenes. In this paper, the evacuation guidance system is defined as a reinforcement learning agent and implemented by the CA-DQN method, which provides a novel approach for the evacuation guidance problem. The experiments demonstrate that the proposed method is superior to the static guidance method, and on par with the manually designed model method.

In the field of intelligent crowd video analysis, the prediction of abnormal events in dense crowds is a well-known and challenging problem. By analysing crowd particle collisions and characteristics of individuals in a crowd to follow the general trend of motion, a purpose-driven lattice Boltzmann model (LBM) is proposed. The collision effect in the proposed method is measured according to the variation in crowd particle numbers in the image nodes; characteristics of the crowd following a general trend are incorporated by adjusting the particle directions. The model predicts dense crowd abnormal events in different intervals through iterations of simultaneous streaming and collision steps. Few initial frames of a video are needed to initialize the proposed model and no training procedure is required. Experimental results show that our purpose-driven LBM performs better than most state-of-the-art methods.

No abstract