Autonomous Surveillance for an Indoor Security Robot

Conventional surveillance for a security robot suffers from severe limitations, perceptual aliasing (e.g., different places/objects can appear identical), occlusion place/object appearance changes between visits), illumination changes, significant viewpoint etc. This paper proposes an autonomous robotic system based on CNN (convolutional neural network) to perform visual perception and control tasks. The aims identify all objects moving in the scene verify whether target is authorized person. includes motion detection module, tracking face detection, recognition module. navigation (path planning obstacle avoidance). empirical validation evaluation metrics, such as model speed, accuracy, precision, recall, ROC (receiver operating characteristic) curve, P-R (precision–recall) F1-score AlexNet, VggNet, GoogLeNet, RMSE (root-mean-square error) value of mapping errors. experimental results showed that average accuracy VggNet under four 0.95, it has best performance unstable factors among three architectures. For building maps real scenes, error 0.222 m.