A RELIABLE AND PRODUCTIVE AUTONOMOUS ROBOTIC SYSTEM
2020 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesis
Abstract [en]
This thesis work focused on increasing both the productivity and dependability of an autonomous service robot. The work was carried out at Volvo Construction Equipment in Eskilstuna, Sweden. The high level of autonomy also required a high level of dependability which is why a fail-stop design is added. Autonomous machines will in the future replace the work of humans but in order for this to happen, high productivity and dependability needs to be achieved. Humans needs to have enough confidence that the autonomous machines are no danger to its surrounding environment but also that there is enough efficiency so that it becomes more or equally efficient as a human. The implementations developed in this thesis work can be transferred to similar autonomous systems to increase productivity and dependability of the system. The productivity was increased by implementing an event based highest priority first CAN-bus scheduler. A mission for moving towards a table and picking up an object was developed and the code was optimized by utilizing all of the CPU cores. The most safety critical wires were identified and the effects of a wire failure were mitigated. Reasonableness checks and watchdog timers were implemented to improve the dependability. Collision detection was implemented to be able to avoid damage to the robot or the environment. Tests showed that the robot performed the mission successfully and is more productive with the optimized code. If the communication between the microcontrollers and the PC was lost then the robot stopped so the fault didn’t propagate. Sensor wire failures or sensor failures could be detected with the reasonableness checks. The robot was able to detect collisions and stop. Some of the implementations and experiments were not done due to the COVID-19 outbreak that caused a lock-down.
Place, publisher, year, edition, pages
2020. , p. 40
Keywords [en]
Robotics, Dependability, Productivity
National Category
Robotics and automation Embedded Systems
Identifiers
URN: urn:nbn:se:mdh:diva-48823OAI: oai:DiVA.org:mdh-48823DiVA, id: diva2:1442075
External cooperation
Volvo Construction Equipment
Subject / course
Computer Science
Presentation
2020-06-05, 14:05 (English)
Supervisors
Examiners
Note
Presentation were done in Zoom due to the COVID-19 breakout.
2020-06-172020-06-162025-02-05Bibliographically approved