XOps IRL: Autonomous Robot - Episode 1

Key Concepts

• Autonomous Robotics
• ROS/ROS2 (Robot Operating System)
• LIDAR (Light Detection and Ranging)
• Edge Computing
• MQTT (Message Queuing Telemetry Transport)
• AI Gateway
• Prompt Injection
• LLM/SLM (Large/Small Language Model)
• Reverse Proxy
• Bill of Materials (BOM)

Project Overview: Autonomous Robot with Security Use Cases

Jason Rom and Joe Martin are collaborating on a project to build an autonomous robot and explore security use cases related to F5 solutions. The goal is to create a fun, accessible project that individuals can replicate at home with minimal cost.

Customer Demands and Requirements

• Customer demands and requirements are the main drivers for use cases.
• Creativity and innovation are needed to meet customer demands.
• Customer requirements are not always natively built into the product.
• Flexibility and programmability are needed to meet customer use cases.

Robot Design and Functionality

• Autonomous Navigation: The robot will be able to roam and navigate its environment autonomously.
• Voice Control: Control the robot via voice commands using a large or small language model (LLM/SLM).
• Security Use Cases: Explore potential vulnerabilities and security risks, such as prompt injection, to make the robot do something it's not intended to do (e.g., enter restricted areas).
• AI Gateway: Potentially use an AI gateway to manage interactions with the LLM and control the robot's behavior.
• Edge vs. Central Processing: Determine the balance between edge processing (on the robot) and central processing (on a hub).

Robot Components and Technologies

• Controller: ESP32 Wi-Fi module will be used as the robot controller.
• Environment Mapping:
  • LIDAR: Use LIDAR for 360-degree environment mapping. Due to budget constraints, it will likely be a single-level LIDAR. LIDAR allows the robot to map, save, and navigate its environment, even detecting new obstacles.
• Mobility: Consider different mobility options like wheels, legs, or treads. Mechanum or omni wheels are preferred for maneuverability.
• Object Detection:
  • Camera Module: Add a camera module for object detection. The camera might be connected via SPI interface or Wi-Fi.
• Communication:
  • MQTT: Potentially use MQTT for communication, with Mosquito as the MQTT broker. Load balancing MQTT can be done through Mosquito, potentially hosted on the edge device.

Software and Operating System

• ROS/ROS2: Utilize ROS (Robot Operating System) or ROS2 for its existing infrastructure.
  • Micro/Mini ROS: Due to the ESP32's limited compute power, use Micro ROS or Mini ROS on the robot.
  • ROS Server: Run the full ROS server on the edge device or a separate machine. ROS typically runs on Ubuntu 22 or 24.
• Edge Device:
  • Raspberry Pi: Initially use a Raspberry Pi as the edge device.
  • K3S/K8s: Potentially containerize applications on the Raspberry Pi using K3S or K8s.

Communication and Security

• Camera Feed: The camera ESP32 can publish a camera image topic, and the edge device can subscribe to it using ROS's native publisher-subscriber architecture. Alternatively, MQTT can be used for streaming the camera feed.
• Engine X: Use Engine X as a reverse proxy to manage connectivity into the environment.
• AI Gateway: Implement an AI gateway to manage interactions with the LLM and prevent prompt injection attacks.

Speech to Text and Intent Conversion

• MCP Tool: Find or build an MCP (presumably "Machine Control Protocol") tool to convert speech to text and then translate the text into commands for the robot.

Budget and Timeline

• Target Budget: Aim for a bill of materials (BOM) cost of around $200.
• Timeline: The project is expected to take 9-12 weeks.

Conclusion

The project aims to build an autonomous robot with a focus on exploring security vulnerabilities and integrating F5 solutions. The team will use a combination of hardware and software components, including LIDAR, ROS, MQTT, and potentially an AI gateway. The project will be documented and shared with the community, allowing others to replicate and contribute. The goal is to keep the project affordable and accessible to hobbyists and enthusiasts.