Theory & Background

Tracked robots are ideal for SAR due to superior obstacle climbing (can surmount obstacles 60–80% of track height), better traction on debris, rubble, and wet surfaces, lower ground pressure distribution, and stability on slopes up to 45°. Configure encoder-based odometry for each track. Drive via skid steering: both tracks forward = straight, one faster = curve, opposite directions = pivot turn.

Components & Requirements

10 components required for this project.

#	Component	Purpose	Qty
1	Tracked Robot Platform (heavy duty)	Rough terrain mobility	x1
2	FLIR Lepton 3.5 Thermal Camera	Human body heat detection	x1
3	MQ-2 Gas Sensor Array	CO, LPG, smoke detection	x3
4	Raspberry Pi 4 (8GB)	Main processing unit	x1
5	4G LTE Module (SIM7600)	Long-range communication	x1
6	RPLidar A2	3D environment mapping	x1
7	Pan-Tilt Camera (HD IP)	Operator view and victim identification	x1
8	IMU (BNO055)	Attitude and heading reference	x1
9	CO2 Monitor (MH-Z19)	Air quality for victim survivability	x1
10	LED Floodlight (12V, 20W)	Illumination in dark rescue environments	x2

Step-by-Step Implementation

Follow these 4 steps carefully.

1

Tracked Platform Assembly for Rough Terrain

Tracked robots are ideal for SAR due to superior obstacle climbing (can surmount obstacles 60–80% of track height), better traction on debris, rubble, and wet surfaces, lower ground pressure distribution, and stability on slopes up to 45°. Configure encoder-based odometry for each track. Drive via skid steering: both tracks forward = straight, one faster = curve, opposite directions = pivot turn.

2

Thermal Camera Integration for Victim Detection

FLIR Lepton 3.5 provides 80×60 pixel thermal image at 8.7Hz via SPI. Each pixel is a temperature reading. Human body temperature (36–37°C) appears significantly warmer than structural wreckage (< 25°C). Implement a threshold detector: find all pixels > 33°C in an otherwise cool scene. Cluster adjacent hot pixels — clusters > 20×20 pixels likely indicate a person. Alert operator with audio alarm and highlight on display.

3

4G Teleoperation System

Stream compressed H.264 video (main camera + thermal overlay) via 4G link. Implement control via UDP socket: joystick commands sent from operator station at up to 50Hz. H.264 video at 720p: 1–3 Mbps, acceptable latency < 200ms on 4G. Use WebRTC for browser-based operator station. Implement control input smoothing to compensate for network latency jitter. Add automatic re-connection on link loss with teleoperation lock (safe stop) on signal loss > 2 seconds.

4

Gas Sensor Fusion for Hazard Assessment

Mount MQ-2 (CO/smoke), MQ-7 (CO only), and MH-Z19 (CO2) sensors. Read every 5 seconds. Display readings: CO < 50ppm (safe), 50–200ppm (caution), >200ppm (evacuate). CO2 > 5000ppm indicates poor ventilation — victim may be unconscious from oxygen depletion. LPG/methane > 1% LEL indicates explosion risk. Log all readings with GPS coordinates to map hazard zones.

Code & Implementation

Core code for sar_robot.py:

sar_robot.py
  Python

import socket, cv2, numpy as np, struct, threading

class SARRobot:
    def __init__(self):
        self.cmd_sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        self.video_cap = cv2.VideoCapture(0)

    def process_thermal(self, thermal_array):
        """Detect potential victims in thermal image"""
        mask = (thermal_array > 33) & (thermal_array < 40)
        contours, _ = cv2.findContours(mask.astype(np.uint8), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
        victims = []
        for cnt in contours:
            if cv2.contourArea(cnt) > 100:  # Min area threshold
                x,y,w,h = cv2.boundingRect(cnt)
                victims.append({"bbox": (x,y,w,h), "temp": np.max(thermal_array[y:y+h, x:x+w])})
        return victims

    def send_motor_command(self, left_speed, right_speed, server_ip):
        cmd = struct.pack('ff', left_speed, right_speed)
        self.cmd_sock.sendto(cmd, (server_ip, 9999))

Testing & Troubleshooting

Test Search and Rescue Robot by verifying each subsystem individually before full integration.

!

Troubleshooting Tips

Verify power voltages, check ground connections, use serial monitor for debug.

Real-World Applications

*Post-earthquake victim search

*Building collapse rescue assistance

*Mine accident rescue support

*Hazardous material incident response

*Flood rescue scouting

*Nuclear facility emergency response

*Forest fire survivor search

*Terrorist incident tactical support

Extensions & Next Steps

Add acoustic sensor for voice/heartbeat detection through rubble
Implement autonomous victim mapping using SLAM + thermal fusion
Add a manipulator arm for light debris removal
Build communication relay capability for trapped survivor communication
Integrate with incident command system software for multi-agency coordination

Interactive Playground

Coming Soon

An interactive simulator will be available here — simulate circuits and run code in-browser without hardware.

Frequently Asked Questions

How does the thermal camera detect victims under rubble?

Thermal cameras detect infrared radiation, which can penetrate thin layers of dust and fabric but not solid concrete or thick rubble. A person under debris: if rubble layer is < 30cm wood/plasterboard, body heat may warm the surface slightly (0.5–1°C above ambient) — detectable with sensitive thermal cameras. Through thick concrete walls: thermal cameras cannot detect. In these cases, acoustic vibration sensors, CO2 sensors (elevated CO2 near trapped victims), or specialized ground-penetrating radar are used.

Introduction