This FTP proposes development of an infrastructure for robot prognostic and health management tool using model-based methodology namely “Prognostics and Health Management tool for ROS”. This tool will be responsible for managing / monitoring robots’ health, RUL (Remaining Useful Life), Probability of Task Completion (PoTC) etc. User is able to enter the necessary equations and component informations (hazard rates, robot configuration etc.) to the PHM tool and the other sensory data like temperature, humidity, pressure, load etc. In addition to this, a PHM package will be developed for the Mobile Robots using this tool.
In the ongoing studies, development continues on the PHM tool for mobile robots.
Studies Done So Far
In this project MS-1 (algorithm development) and MS-2 (tool/GUI development) are completed successfully. The algorithms developed in MS1 have been implemented in to the tool which is developed in MS2. In MS-2, ROS packages and Human-Machine Interfaces (HMI) is developed for developed algorithms in the first milestone. Besides, ROS topics, required message types are generated for the nodes. HMI is developed by using QT and contains customized interfaces like user robot configuration setup sub-tool, analysis sub-tool, algorithm selection / implementation sub-tool, create formula sub-tool, reliability analysis sub-tool, anomaly detection sub-tool, health monitoring sub-tool etc. Besides, user could generate special interfaces by using RQT and other developed sub-tools. The other ROS packages could be implemented as generic packages for inputs from user, pre-defined configuration or algorithms. While doing these developments on the tool, it has been paid attention that the developed tool (GUI) is user-friendly.
By using tabs for each process to be performed in the developed interface, the user is enabled to work more regularly. For instance, if the user wants to calculate the failure rate of a battery, he should select the mechanical equipments tab. Or, he should select the electrical equipments tab if he wants to calculate the failure rate of a capacitor.
– Fig 1. Calculation of the capacitor’s failure rate –
– Fig 2. Calculation of the battery’s failure rate –
While doing the mathematical calculations some data types are used such as ambient temperature, pressure, current on the system. Therefore, calculations are updated simultaneously with the change of this data. Thanks to this developed tool, changes in calculations (failure rate, reliability etc.) are displayed graphically to the user. The graphics are displayed by clicking the “View Graph” button.
– Fig 3. Capacitor failure rate change over time-
In this tool, the user can create a custom system while doing these things:
- The user can add custom components into the system.
- The user can generate custom modules by using custom and default components.
- The user can generate and load his module specific formula.
By doing this, the user can have the freedom to develop the system in any way he wishes. In the following figures, the current situation in the interface used to customize the system is shown.
– Fig 4. The tab of custom modules and formulas are created –
– Fig 5. The tab of the custom components are created –
In the “Set Type” tab, the user can arrange the types of the components under the modules and arrange the types of the modules under the system. The types are arranged as serial or parallel. In this way, the reliability value of the system is calculated by adjusting the types.
NOTE: Algorithms for use in the “Set Type” tab are currently under development. Accordingly, this image does not represent its final form.
– Fig 6. The tab of types are setting –
In the “Monitoring” tab, the hazard rate value, reliability value and probability of task completion (POTC) of the system are shown to the user, upon the given data. Also, their graphical representations are shown. The biggest novelty in this system is these values can be observed in real time.
– Fig 7. The Monitoring tab of Hazard Rate –
– Fig 8. The Monitoring tab of Reliability –
While calculating the system reliability and POTC, tasks are executed using the mobile robot in Gazebo.
– Fig 9. – Fig 9. The Monitoring tab of POTC – The Monitoring tab of POTC –
– Fig 10. The tasks are executed using the mobile robot in a Gazebo environment –
– Fig 11. State diagram of tasks is shown using Smach –
– Fig 12. Robot’s path and obstacles in Rviz.-
In this study, Milestone 1 which is algorithm development and Milestone 2 which is tool/GUI development is completed successfully. The other milestone which is MS-3 PHM package for Mobile Robots are under development.
In MS-3 PHM package for Mobile Robots will be developed to show the effectiveness of PHM tool. Mobile robots are selected because we have expertise on this area. PHM package for Mobile robots will contain nodes, launch files, formulas, special robot configuration file, rosbag files for sensory information of mobile robot.
For the details: