AstaZero: Orchestrated Control of Drones

Overview

AstaZero tasked this team to collaborate with students at Chalmers University to automate the filming of EuroNCAP tests conducted by AstaZero. Currently, the filming for these tests is done manually which results in wasted time and labor. To solve this problem, AstaZero wants to automate this filming process using drones. The goal of this project is to write software that will allow the drones to film the EuroNCAP tests for AstaZero at optimal angles without active user input.

Objectives

For this project, the main objective is to automate the filming of vehicles undergoing the EuroNCAP tests using drones. The positioning of the drones must be in accordance with the standards outlined in the EuroNCAP tests. This objective was met by developing an app that acts as interfacing software between the drones and their AV-Test Operating System (ATOS), a test control system used in executing complex traffic scenarios. Furthermore, PSU was given the task of creating an object detection module that could send alerts for unwanted objects detected on the test track. The team at Pennsylvania State University was in charge of converting ATOS commands into trajectories to be sent to the drone and developing a separate object detection module. Furthermore, the students at the Chalmers University of Technology focused on the object-detection algorithms for the drone to ensure that the drone is following the car.

Approach

For the Penn State team, development was divided into the trajectory module and object detection module. The First goal was accomplished by developing functions outside of the main ATOS system. The path file for a test object was parsed into usable data to create a new path object for the drone to follow that would be used by the ATOS system. After the functionality was accomplished, the module was incorporated into ATOS for use in the EuroNCAP system. AstaZero identified a need for creating an alert system for use during AstaZero’s testing. The goal of this feature would be that if any unidentified objects entered the camera view during testing, an alert would be sent to the AstaZero employees on the test track.

Outcomes

Due to many roadblocks and limited time and resources, our final product was constrained to a foundational building block for potential further development. The Penn State team did not have access to the drone, a lot of software development was still being finished by AstaZero, and the scope of the project changed numerous times. Due to this, the trajectory module was limited to a specific filming scenario and testing was not accomplished. Furthermore, the same issues plagued the development of the alert system module. The module is close but not completely finished, and still requires future testing. Overall, we accomplished and learned a great deal, but there is still room for further development within this project.