| Aiming at the problems of uncontrollability,unpredictability and non-reproducibility in the research of rat's behavior,which is urgently need to be solved for the biologists,an intelligent interactive control system was proposed.Different behaviors(stressful,neutral,friendly),which aims to stimulate a laboratory rat were generated by controlling a self-designed biomimetic robotic rat with binocular vision.The response of a laboratory rat to a biomimetic robotic rat was studied.According to the multi-degree of freedom,the wheel-legged structure of the biomimetic robotic rat and the driving characteristics of different motors,the behavior models(moving,rearing,mounting,rotating and grooming etc)are established.The "improved Camshift algorithm based on Kalman filtering" was introduced to realize real-time tracking and positioning of the target laboratory rat.This paper also designs "a capability of behavior cognition for the biomimetic robotic rat based on SVM and ANN" to achieve behavior recognition and behavioral classification.Based on binocular vision system,this system can effectively obtain the three-dimensional information of rats.Combined with an efficient image classification processing algorithm,the system can accurately analyze the experimental rat's behavioral characteristics,and modulate the behavior of rats by controlling the behavior of the biomimetic robotic rat,so as to achieve intelligent interaction.The social interaction experiments between the laboratory rats and the biomimetic robotic rat show that the system is able to capture rat's behavioral characteristics in real time and direct rat's behavior by controlling the behavior of robotic rats.The system provides biologists,neurologists and brain scientists with a research platform and an assistant research tool.Furthermore,this research has potential application value in the treatment of mental disorder and screening of psychotropic drugs in the field of psychology and pharmacology. |
PDF Full Text Request