As one of the founding members of the Autoware Foundation, it is our mission to support the open source community, and ensure that we continue to accelerate the safe development and deployment of self-driving technology.
As such, we are releasing a tool which will allow developers to test and validate their AI software stacks; StreetDrone’s Vehicle Simulation model of the SD Twizy, is now available for public use here: https://github.com/streetdrone-home/SD-TwizyModel
Try the model out with our suggested Gazebo worlds and visualize your sensors’ output on the RViz visualization tool. The model supports 8 x Sekonix Color cameras and the Velodyne VLP-16 LiDAR as main localisers. We suggest controlling the robot using a Joystick (more info can be found on our README files) and navigate around the worlds observing the generated LiDAR point cloud and the cameras’ raw images output in RViz.
With the help of Oxford Brookes University, we retrieved the SD Twizy’s vehicle dynamics measurements on their 4 post rig, and built a detailed simulation model as representative and realistic as possible.
This release represents an initial starting point, and should give every aspiring developer, a shortcut to getting their software stacks up and running. Within our repository you will find detailed instructions on how to set up the environment and run the simulation, as well as how to contribute to this project and report an issue.
Should you require any help on setting the project on your local machine, or would like to discuss how we can support moving from software simulation to real world testing, then please do get in touch.
Here’s a video of Gazebo, RViz and our simulated Twizy in action:
Become a Member
Enter our draw for a chance to test drive the StreetDrone ONE and spend a day with the SD team! Additionally, get acces to exclusive technical content, discounted events and conferences