Hey, we are ARTIsts

We are searching for an Robotic Software Developer (30-38,5h/week)

Tasks:

• Software development for autonomous robots
• Optimize and analyze existing software solutions
  
• Evaluate the software in simulations and real world prototypes
  
• Supervise projects independently and manage customers contacts

Requirements:

• Education in the field of software development
• Experience in C, C++, Python
• Experience with mobile robotics and ROS (Robot Operating System)
• Willing to travel to European countries
• Work independently and with a hands–on mentality
• Good English skills

Automatic configuration of a base driver for robots. Using a description of the robot (e.g. by a URDF), kinematic parameters are derived (e.g. wheel distance, etc.).

These are then used to configure the base driver. Automatic tests are then carried out to recalibrate the values obtained. This is to reduce the deviation between “should” and “is” (keyword: manufacturing tolerances).

Starting from an initial map created with a LiDAR-Scanner (2D as most basic variant), the original map is to be successively improved and better understood by incorporating changes in the environment over time.

New map information data will be sent to an intermediary server, where data is processed with an existing ARTI-AI Kit (LOKI). When data is processed on server-level, the new map-information is compared with existing mapping-material. If the comparison finds significant changes, a new map is created. This map is then fed back to the robot.

The research questions to be worked on are: finding parameters to define, whether changes in the map are “significant” or not and to develop a pipeline to fed the newly updated maps back to the robots.

The starting point is a “wizard” that is already being developed to facilitate the integration of the ARTI-AI software kits in customer vehicles.

By means of simple tests/tasks, basic configurations; first integration steps should be carried out also by non-robotics experts.

Within the scope of the scientific work, it is to investigate which parameters must be automatically derived by the wizard-application to ensure a functioning and reliable navigation when the starting point is an already precisely localized and calibrated robot.

In developing software for autonomous mobile robots, testing plays an important role. The software can, to a certain extent, be tested virtually. But at some point, the tests must be performed in “real world scenarios”. Such tests can be quite repetitive and time-consuming. Within a master’s thesis framework, we want to explore how such test scenarios can be carried out faster and more automatically.

As a first task, the thesis should analyze the current testing workflow and estimate how much time can be saved by integrating the proposed testing methods. Throughout the thesis, different testing methods should be explored to allow fast and accurate testing of different types of ground-bound robotic vehicles. The focus should be on strategies that automatically execute the tests for each release.

The thesis should cover different levels of testing, from unit tests up to integration tests, including showcases for each level and every part of the system. To ensure proper execution on the Microcontroller, at least the integration tests should be executed (also) on the Microcontroller.

Additionally, the thesis should cover the tooling needed to allow the fast creation of additional tests apart from the showcases. Apart from defining proper test methods, the thesis should also cover issues of appropriate test coverage to avoid creating too few/too many test cases.