Robotics Projects in Bangalore for ECE

A differential-drive robot running ROS Noetic on a Raspberry Pi 4B uses a RPLIDAR A1 for simultaneous localisation and mapping (SLAM) via the gmapping package. The robot navigates autonomously in a 5×5 m indoor environment using the ROS Navigation Stack with AMCL localisation, a global planner (Dijkstra), and a local planner (DWA). The robot avoids dynamic obstacles in real time and reaches goal poses with a position accuracy of ±5 cm. A RViz visualisation displays the live map, robot pose, and planned path. The project demonstrates ROS architecture, SLAM, and autonomous navigation.

A 4-degree-of-freedom robotic arm built from MG996R servo motors and 3D-printed links is controlled by an Arduino Mega running an inverse kinematics solver based on the geometric method. The arm picks objects from a fixed source position and places them at one of four target positions identified by colour using an OV7670 camera and a colour blob detection algorithm. A PS2 joystick provides manual override control. The IK solver computes joint angles in under 5 ms, enabling smooth 0.5 Hz pick-and-place cycles. The project demonstrates robotic kinematics, servo control, and basic machine vision.

A swarm of 20 virtual robots is simulated in Python using the Pygame library, implementing Reynolds' flocking rules (separation, alignment, cohesion) with an added goal-seeking behaviour. Each robot has a sensing radius of 50 pixels and updates its velocity at 30 Hz. The simulation demonstrates emergent collective navigation through a maze with static obstacles, with 95% of robots reaching the goal within 60 seconds. Parameters (sensing radius, weight coefficients) are adjustable via a GUI slider panel. The project demonstrates swarm intelligence principles, multi-agent simulation, and emergent behaviour analysis.

An 18-servo hexapod robot is controlled by two Arduino Mega boards communicating over I2C. The primary board runs a tripod gait controller that computes servo angles for all six legs using a simplified inverse kinematics model, achieving a walking speed of 15 cm/s on flat surfaces. An HC-SR04 ultrasonic sensor on the front triggers an obstacle avoidance manoeuvre. Gait parameters (step height, stride length, speed) are adjustable via a Bluetooth serial interface from a smartphone app. The project demonstrates multi-servo coordination, gait planning, and real-time embedded control.

A wearable glove equipped with two MPU6050 IMU sensors on the forearm and hand transmits orientation data over Bluetooth to an Arduino Mega controlling a 5-DOF robotic arm. Pitch and roll angles from the forearm IMU map to shoulder and elbow joints; wrist flexion from the hand IMU maps to the wrist servo. A flex sensor on the index finger controls the gripper. The system achieves a control latency of 80 ms from gesture to arm movement. The project demonstrates IMU-based gesture recognition, Bluetooth serial communication, and servo control for teleoperation.

A wheeled robot equipped with a Raspberry Pi 4B and Pi Camera Module 3 traverses crop rows and detects weeds using a MobileNetV2 classifier fine-tuned on a custom dataset of 2,000 images (crop vs. weed). Detected weed locations are marked on a grid map and a servo-actuated sprayer nozzle dispenses herbicide only at weed positions, reducing chemical usage by an estimated 60% compared to broadcast spraying. The robot navigates using line-following sensors along crop row markers. The project demonstrates precision agriculture, on-device deep learning inference, and robotic actuation.

An Arduino Uno-based fire-fighting robot uses three flame sensors (left, centre, right) to detect and navigate toward a fire source. A PID-like steering algorithm centres the robot on the strongest flame signal. When within 20 cm of the flame, a 5V mini water pump activates for 3 seconds to extinguish it. An ultrasonic sensor prevents the robot from colliding with walls during navigation. The robot successfully extinguishes a candle flame in a 1×1 m arena in under 15 seconds in 90% of trials. The project demonstrates sensor-guided navigation and reactive robot control.

An underwater remotely operated vehicle (ROV) is built using a waterproofed acrylic frame with four bilge pump motors as thrusters, controlled by an Arduino Mega via ESCs. A Raspberry Pi Zero W streams a live video feed from a waterproofed USB camera over Wi-Fi to a surface laptop. Depth is maintained using a MS5837 pressure sensor and a PID controller adjusting vertical thruster thrust. The ROV is tethered with a 5-metre cable carrying power and Ethernet. The project demonstrates underwater robotics, thruster control, pressure-based depth sensing, and live video streaming.