Armed forces and civilian securities helps in promoting and maintaining a peaceful harmony in our country and the civilian industries to ensure safety. The risk and complications makes the job more tough and stressful. A survey on over 70 percent of our forces says, they were not getting adequate rest and sleep. In Order to ease their task and automate the security functions, ASTRA (Attack Suspecting and Tactically Responding Army machines) is developed.
ASTRA is network of swarm UGV/UAV robots collaborated together for surveillance, targeting, and improved communications. It uses modern technologies like computer vision, AI, Encrypted Communication and sensor based interactions. It is provided with algorithms that helps the machines to perform mostly like humans which includes real-time communication,vision-based target tracking ,alert the securities about intruders and unknown suspects, which can reduce the occupational stress on soldiers.
ASTRA can be deployed at any area in need of Surveillance. The real-time camera feedback and Automated Exploration feature (between two GPS coordinates) ease the task of Surveillance. The Manual/Remote controlled exploration feature also makes the traversal easier. The UGV can detect the unusual motions and alert the base Security station. It can track the object and follow it till the security force covers the area of suspect. The alert is also passed to the inter-connected robots which makes the robots more vigilant. By this, even if the tracking robot gets destroyed, the other one will get into duty of the previous one(follow the suspect). Thus, it can ensure more security.
The UGV is build on an all wheel drive aluminium chasis. The drive train consists of 4 high torque brushed DC motors with planetary gearbox. The battery is made from 18650 Lithium ion cells with an overall power rating of 40Ah at 13.2 volts. The battery is protected using battery management system which protects it from over charding and discharging. DC-DC step down converters are used for powering the sensors are onboard computers. The navigation system of the robot consists of laser scanners, ultrasonic sensors, GPS, IMU and camera for obstacle avoidance, localization and mapping. The data processing is done in Atmega2560 MCU which then publishes this data to a Raspberry Pi running Ubuntu. The ROS drivers for the ATmega MCU publishes all the sensor data to the Raspberry Pi and subscribes velocity commands from it. SLAM algorithm is used for mapping and localization of unknown environments. The computed map is then used for patrolling around the environment. Raspberry Pi camera is used for detecting the presence of intruders and uses machine learning for improved detection even at low light conditions. The robot has high intensity IR led array which can improve the visblity upto 25 m at dark environments. All the robots are connected to a central PC running ROS master node via the onboard Wifi. The prototype was tested in an an area which is free from other radio interferences and the distance between the robots were maintained at 5 m for low latency communication.