| Issue number | Date found | Date fixed | Made by | Description | Resolution |
|---|---|---|---|---|---|
| 17 | 11/27/2019 | 12/06/2019 | Harsh Sharma | Simulation computer does not have updated CMake, updating it would break ROS near FVD | Installed second instance of CMake |
| 16 | 11/15/2019 | 11/27/2019 | Sohil Savla | Images were published at 2Hz | Switched to Webots ROS controller that publishes images by itself. |
| 15 | 11/25/2019 | 12/07/2019 | Sohil Savla | 3D point cloud is generated a 2 Hz | Switched to C++ implementation of PCL code. |
| 14 | 11/29/2019 | 12/06/2019 | Ayush Jhalani | State Machine overloaded with continuous subscriber creation | Switched to rospy wait_for_message() |
| 13 | 11/27/2019 | 12/06/2019 | Sohil Savla | Local Planner fails at other locations | An accurate costmap is generated by studying simulation modelling of 2.5D map |
| 12 | 11/15/2019 | 11/27/2019 | Ayush Jhalani | Local Planner fails due to incorrect costmap | Added an inflation radius around the cost-map to alleviate issues where robot falling into pit |
| 11 | 11/15/2019 | 11/19/2019 | Himil Sheth | Unable to align frames for multiple parts: planner, RVIZ and Webots | Lucky break in aligning one set of frames helped uncover the confusion. |
| 10 | 10/20/2019 | 11/12/2019 | Sohil Savla | Custom rocks created tend to sink into the surface | Changing the density of the rock changes the sinking properties |
| 9 | 04/22/2019 | 05/14/2019 | Harsh Sharma | Almost always, the camera throws a host of varied errors after clicking 3-4 photos. Meanwhile, the camera works perfecty in some instances. | The problem causing this issue has not been recognized completely yet. |
| 8 | 04/20/2019 | 05/10/2019 | Harsh Sharma | The present method of detection on the rovers is a color based one. After conducting various tests and experiements, the concusion was that this method is highly unreliable. It is highly environment and lighting condition dependent. The exposure and focus had to be tuned in every single trial. | A more robust method of detection is being worked on. This includes training a neural net to detect the rover with high accuracy. It would be a move away from color based detection. |
| 7 | 04/22/2019 | 04/23/2019 | Harsh Sharma | The RTK GPS and the UWB sensor reading had to be compared, however they were mounted on different sides of the same level and hence there were discrepencies when the two readings were compared. | The quick fix to this problem was to change the location of the UWB sensor, to somewhere really close to the RTK antenna. This was enough for our requirements. However, the permanent and a more accurate fix to this would be to compute the transformations between the two devices. |
| 6 | 04/22/2019 | 04/23/2019 | Harsh Sharma | The initial placement of the RTK GPS was facing a lot of interferences. Its occluded neighbourhood was not letting it receive appropriate satellite signals. The placement of the RTK GPS also interferred with the point grey camera which was on the same level as that of the RTK GPS antenna. | The RTK GPS antenna was mounted on a meter long rod raising it above all the existing components. This reduced/removed the interferences and appropriate satellite signals were then able to be captured. |
| 5 | 05/01/2019 | 05/20/2019 | Harsh Sharma | The day of our SVD Encore, the pinion sheared off even after the installation of the limit switch to limit the rotation of the same. | Replacement of the sheared pinions and also limiting the maximum current limit for the steering motor. |
| 4 | 04/20/2019 | 05/20/2019 | Harsh Sharma | The WiFi was experiencing latency issues during the trial tests. It also seemed to automaticaly set to some other IP configuration than the one it was configured to. | The WiFi was tuned to operate on 5GHz network since that experiences less traffic at the CMU campus. The Wifi failed to opearate at that range and we reverted back to 2.4GHz. The permanent solution to the problem of setting of IP to random configurations would be to make the router static in the network using the MAC Address. |
| 3 | 02/28/2019 | 03/04/2019 | Sohil Savla | Slipping on project timeline due to fatigue from the previous test | Have regular team meetings and standups to ensure proper tracking of tasks |
| 2 | 01/29/2019 | 04/27/2019 | Sohil Savla | Steering Rack and Pinon system is unreliable and takes very long to fix | Redesign Steering mechanism |
| 1 | 01/10/2019 | 01/11/2019 | Sohil Savla | Steering Rack and Pinon system fails | Change Pinon |
