Abstract
In the aftermath of natural disasters or terrorist activities, the "golden hour" for rescue is extremely limited. This paper presents the design of an autonomous intelligent robot controlled by an S3C2410 MCU. The system utilizes a wireless network card and UDP protocol to receive commands and transmit real-time video, environmental data (temperature, humidity, methane), and GPS coordinates to a host system, facilitating efficient search and rescue operations in hazardous ruins.
Introduction
Disaster environments, such as collapsed buildings or mines, are often inaccessible or too dangerous for human rescuers. While traditional robots exist, wheeled variants struggle with debris, and legged variants are often slow. This research addresses the need for a versatile, hybrid locomotion system that combines the speed of wheels with the obstacle-clearing capabilities of a six-legged bionic structure, modeled after the anatomy of an ant.
System Architecture
The system is divided into two main components: the Walk Detection Robot and the Control Display Terminal, communicating via a wireless LAN.
Hardware Design
The robot's "brain" is an S3C2410 (ARM9) processor running at 203MHz, which manages high-level data processing and network communication. To ensure precise movement, a secondary ATmega16 (AVR) microcontroller is used as a dedicated servo controller, handling 11 channels of PWM signals for the robot's legs and steering. The sensor suite includes a USB camera for visual feedback, sensors for temperature, humidity, and methane, and a GPS module for localization.
Software and Communication
The robot runs on Linux 2.6.24. The control interface is developed using the Qt (QTE) framework. For communication, the UDP protocol was selected over TCP to prioritize low-latency video transmission, essential for real-time remote navigation.
Gait Control Strategy
The robot employs a "composite mode" locomotion strategy, switching between wheeled and legged movement based on the terrain's complexity.
The control logic evaluates obstacle height ($\Delta v$) against specific thresholds:
- Wheeled Mode: Used when ground irregularities are minimal ($\Delta v < \Delta y$).
- Climbing Mode: Activated when obstacles are within the leg's reach ($\Delta y < \Delta v < \Delta x$).
- Avoidance Mode: Triggered if an obstacle is too high to climb ($\Delta v > \Delta x$).
Key Results
Functional testing demonstrated the system's viability for disaster relief:
- Real-time Video: The system achieved a stable transmission of 20 frames per second at a resolution of 320x240.
- Network Performance: The 54Mbps WiFi link successfully handled concurrent streams of video and sensor data.
- Locomotion: The robot successfully transitioned between modes to navigate obstacles, with control latency remaining within acceptable limits for remote operation.
Conclusion and Future Work
The hybrid wheel-leg design provides a balanced solution for navigating unstructured disaster environments. Future developments will focus on improving signal robustness in concrete-heavy ruins using WiFi relay modules, developing a formal kinematic model for optimized movement, and increasing image resolution and battery endurance.
Summary updated