Search for a command to run...
A wheeled surveillance robot augmented with artificial intelligence was built and tested for use in laboratory settings where unattended monitoring and hazard detection are required. Three distinct operating modes were incorporated: manual joystick-driven control via a web browser, fully autonomous obstacle-reactive navigation, and a pre-programmed mappedpatrol sequence that traces a fixed route through the monitored space. Motor actuation and sensor interfacing are handled by an ESP8266 (NodeMCU) module, which also hosts a lightweight private web dashboard over Wi-Fi. Temperature, humidity, and air-quality readings are gathered by DHT11 and MQ135 sensors respectively, while an HC-SR04 ultrasonic transducer supplies proximity data for collision avoidance. A smartphone secured to the chassis runs IP-Webcam and forwards a live video stream over the local network to a Flask/OpenCV server running on a separate machine; this server performs motion detection, face-based attendance recording using Local Binary Patterns Histograms (LBPH), and conversational inventory assistance through a rule-based chatbot. Onboard power is supplied by a custom pack built from fifteen recycled 18650 lithium-ion cells wired in a 5P×3S arrangement, protected by a 3S BMS and regulated to 5V via a 7805 linear regulator. Testing inside a 6m × 4m indoor laboratory yielded 88–93% face recognition accuracy, 93% motion-detection precision, and a 90% patrol completion rate across ten runs. The outcome is a low-cost, modular platform that validates the hybrid embedded-plus-server approach for intelligent laboratory oversight.