Challenge
The manufacturer wanted to improve the quality of its robotic lawnmower software and reliably verify the mowers’ positioning, which works without GPS. To this end, the automated testing of the robots was to be modernised.
Solution
Novotec expanded the test environment: automated tests for every change to the main branch using a simulated mower, test rigs for basic functions such as power-up and power-down, a reference system with high-precision satellite positioning and a test setup for indoor use. Development was done in Python and C.
Result
Every change to the software is automatically tested against a simulated mower and on test rigs. The robots’ positioning can be measured against a reference system and evaluated graphically.
Background
The client manufactures garden equipment and offers its customers web services that it develops and operates itself. Its robotic lawnmowers determine their position without GPS. Whether they do so accurately enough can only be checked with an independent measurement.
For this, Novotec built a reference system with high-precision satellite positioning (RTK). The position determined by the robot itself is compared with this reference value. In addition came performance tests of the mowers, a test setup for indoor use and test rigs on which basic functions are tested automatically.
Work was agile, in three-week sprints. In addition to development, this included maintaining the test environment, evaluating the results and supporting bug fixing and incident management.
Our services
- Automated testing of the main branch with a simulated mower (Jenkins, Bamboo, test rigs)
- Automated basic tests of the mowers on the test rig, such as power-up and power-down
- Building a reference system with high-precision satellite positioning (RTK)
- Performance tests of the mowers against the reference system, test setup for indoor use
- Maintaining the positioning system, GPS tests and modifying mowers for testing
- Implementation of the interfaces to GPS and gateways
- Reworking the Python logger: capture and analysis of data traffic separated
- Support in migrating the mower simulation from Windows to Linux
- Graphical evaluation of measurement data, reporting and assessment of test results
- Support in bug fixing and incident management, advice on software engineering
Technology used
Python and C on Embedded Linux; Jenkins and Bamboo; Selenium for the web services; test rigs, simulation and an RTK reference system. Scrum as the working method.