Work at IVI: Developing an Autonomous Agricultural Vehicle Prototype
Introduction:
-
Brief Overview:
My 8-month internship at l’Institut du Véhicule Innovant immersed me in the ARION team, dedicated to advancing Quebecois companies in autonomous vehicle technology. It marked my debut applying autonomous navigation and robot operating system skills in a professional setting, focusing on developing a prototype for an autonomous agricultural vehicle.
Key Details:
-
Project Description:
-
Objective:
The primary objective was to develop a prototype for an autonomous agricultural vehicle capable of navigating predefined GPS paths while avoiding obstacles. This ambitious goal required a multidisciplinary approach, combining elements of robotics, navigation systems, and real-world testing in agricultural environments.
-
Challenges:
- One of the unexpected challenge was working in the field with computers. The visibility on the screen was minimal and the network needed to be extended on great distance wirelessly. To solve the visibility problem, we installed a box trailer with a big open door on one side to see the prototype in the field. And to allow constant communication with the prototype, we installed multiple directional wifi antennas.
- Fine tuning the simulation to get closer to reality results was a challenge.
-
Achievements:
- Integration of RTK GPS: Integration of an RTK GPS into the navigation system provided precise localization essential for navigating agricultural fields with high accuracy.
- Integration of CAN bus with ROS: Integration of a CAN bus with ROS enabled seamless communication between various vehicle components, enhancing overall system reliability and performance.
- Development of Custom Ackermann Steering Controller: Development of a custom Ackermann steering controller optimized vehicle manoeuvrability and control, essential for navigating tight spaces within agricultural settings.
- Enhancement of Simulation Accuracy: Enhancement of simulation accuracy through meticulous fine-tuning brought simulated results closer to real-world conditions, facilitating more effective testing and validation processes.
- Merging Multiple Sensors for SLAM: Implemented the merging of multiple sensors to achieve Simultaneous Localization and Mapping (SLAM). This approach allowed the vehicle to build a map of its environment in real-time while accurately determining its location within the map.
-
-
Skills Demonstrated:
-
Technical Skills:
My role demanded proficiency in a wide range of technical domains, including programming languages such as C++ and Python, familiarity with autonomous vehicle systems, ROS middleware, SLAM algorithms, control system design, version control with Git, URDF modeling, Linux-based development, LiDAR sensor integration, GPS navigation, robotics principles, software design patterns, simulation environments like Gazebo, and network protocols for seamless communication between vehicle components. #skill/hard/Cpppythonautonomous_vehicleROSSLAMcontrol_systemgitURDFlinuxlidarGPSroboticssoftware_patternGazebo_simulatornetworkNvidia_Jetsonlarge_codebase_management
-
Soft Skills:
In addition to technical expertise, successful completion of the project relied heavily on soft skills such as conducting thorough research to identify optimal solutions, effective teamwork and collaboration with colleagues from diverse backgrounds, meticulous documentation of project progress and findings, proficiency in troubleshooting and conducting in-the-field repairs, clear and concise communication with clients to understand and address their requirements, and adeptness at gathering and translating user needs into technical specifications. #skill/soft/researchteamworkdocumentationin_the_field_repaircommunication_with_clientrequirements
-
-
Impact and Results:
-
Quantifiable Results:
The culmination of our efforts resulted in the successful delivery of a fully autonomous grass cutter vehicle to our client. Equipped with GPS guidance, 3D vision capabilities, multiple LiDAR sensors for obstacle detection, and a smart bumper for enhanced safety, the vehicle represented a significant leap forward in autonomous technology for agricultural applications. Its deployment promised increased efficiency and productivity while ensuring the safety of operators and bystanders alike.
-
Visual Elements:
-
Images/Visuals:
N/A