| A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z | AA | AB | AC | AD | AE | AF | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | Risk ID | Jira Item | Title | Req | Owner | Date Submitted | Date Updated | Description | Impact | Programmatic | Technical | Schedule | Finance | Consequence | Likelihood | Severity | Status | Mitigation Strategy | Expected Outcome | Response Strategy | Type | Type | |||||||||||
2 | 1 | Too Many Bugs | All | Dec 1, 2020 | Dec 1, 2020 | Too many resources are devoted to resolving bugs | - Development process is significantly slowed. - Possibly miss deadline/milestone | 4 | 5 | HIGH | Open | - Use Version Control - Do Continuous Integration - Automate test cases | Bugs are easier to identify and resolve | - Track bugs in Jira - Every bug is assigned an owner - Commits are linked to Jira issues | Programmatic Technical Schedule | Programmatic Technical Schedule | |||||||||||||||||
3 | 2 | Interface Mismatch | All | Dec 1, 2020 | Dec 1, 2020 | Subsystem's have incompatible interfaces | Subsystem's Can't Integrate | 5 | 2 | MEDIUM | Open | - Define interfaces for each subsystem - Detailed cyberphysical architecture - Develop from top down | Subsystems can interface with each other | Technical | Technical | ||||||||||||||||||
4 | 3 | Unintelligible Codebase | All | Dec 1, 2020 | Dec 1, 2020 | Codebase is difficult to comprehend | - Team has trouble integrating subsystems - Difficulty resolving bugs | 3 | 3 | MEDIUM | Open | - Document code well - Follow google style guide | Code is easier to comprehend | Technical | Technical | ||||||||||||||||||
5 | 4 | Interface Change | All | Dec 1, 2020 | Dec 1, 2020 | A subsystem changes it's interface | Subsystem's cannot communicate | 4 | 3 | MEDIUM | Open | - Breaking changes are resolved on a seperate branch | - Develop branch avoids known breaking changes - Breaking changes are resolved in isolation | Technical | Technical | ||||||||||||||||||
6 | 5 | It Worked Last Night | All | Dec 1, 2020 | Dec 1, 2020 | The code was working, but we broke it right before the demonstration | No demonstration | 5 | 2 | MEDIUM | Open | - Use version control so we can undo mistakes - Keep seperate prod., dev., and demo branches | We we can undo our mistakes and still have a demonstration | - Roll back to a previous commit | Technical | Technical | |||||||||||||||||
7 | 6 | Covid Lockdown | All | Dec 1, 2020 | Dec 1, 2020 | CMU shuts down beccause of covid | No access to facilities | 2 | 3 | LOW | Resolved | - A teammate agrees to house hardware | We can still develop and validate during a lockdown | - Move hardware to teammates abode | Programmatic | Programmatic | |||||||||||||||||
8 | 7 | It Works On My Machine | All | Dec 1, 2020 | Dec 1, 2020 | Different environments cause compatibility issues | Code works on some machines but not others | 3 | 4 | MEDIUM | Resolved | - Containerize our application | All code runs in the same environment | Technical | Technical | ||||||||||||||||||
9 | 8 | CR-106 | No Forklift | MP 4 | Jinkun, Tom | Dec 1, 2020 | Dec 10, 2020 | Insufficient access to a real forklift | Classification subsystem has not object to use for development or validation | 5 | 4 | HIGH | Resolved | - Buy a toy forklift (12/18/2020) - Request permission to videotape a forklift from local businesses/NREC (Won't Do) - Create a forklift mockup (Won't Do) | We can test and validate our detection and classification system | Programmatic | Programmatic | ||||||||||||||||
10 | 9 | Pushed Broken Code | All | Dec 1, 2020 | Feb 12, 2021 | Teammate pushed code that is broken | Broken codeabase | 3 | 5 | HIGH | Resolved | - Use Version Control - Automatically reject any push that fails a test case (Won't Do 2-12-21) - Rebase before pushing - Use pull requests for large merges into dev. branch | Broken code is likely to be rejected | Technical | Technical | ||||||||||||||||||
11 | 10 | Man Down | All | Dec 1, 2020 | Dec 10, 2020 | Teammate gets covid or otherwise cannot work | Lost man hours, emotional and physical trauma, death | 5 | 3 | HIGH | Open | - Everything has backup person assigned - Document work - Discuss project priorities with OMRON (Feb 19) | - Other teammates can take over work - Requirements are relaxed in proportion to lost man hours | Programmatic Schedule | Programmatic Schedule | ||||||||||||||||||
12 | 11 | Laptop Down | All | Dec 1, 2020 | Dec 1, 2020 | Laptop fails | Cannot work or code on a laptop | 4 | 2 | MEDIUM | Open | - Members of the team are able to demo from their laptop - Containerize application | - Demos can be done as long as there is 1 working laptop - It is easy to get started on a new machine | Programmatic | Programmatic | ||||||||||||||||||
13 | 12 | Hardware unavailable | All | Dec 1, 2020 | Dec 1, 2020 | Seller of commercially-available hardware/robot platform no longer able to sell (e.g. due to COVID reasons, out of stock, etc.) | Need to search for other available hardware platform to validate on | 4 | 2 | MEDIUM | Resolved | - Research multiple options for hardware platforms (Dec 18) - Order as early as possible (Dec 18) | Technical Schedule | Technical Schedule | |||||||||||||||||||
14 | 13 | Sponsor Unsponors Us | All | Dec 1, 2020 | Dec 1, 2020 | Sponsor stops sponsoring us | We no longer have a sponsor | 5 | 1 | MEDIUM | Won't Do | - Find alternative sponsor | Programmatic | Programmatic | |||||||||||||||||||
15 | 14 | Need More Data | MP 4 | Jinkun | Dec 1, 2020 | Dec 1, 2020 | We cannot find or generate a sufficient dataset for forklift classification | Classifier NN cannot be trained on a forklift | 5 | 5 | HIGH | Resolved | - We create a mockup and make our own training dataset | We can detect and classify our mockup forklift | Technical | Technical | |||||||||||||||||
16 | 15 | No time to do all avoidance algorithms | MP 5 | Tom | Dec 1, 2020 | Dec 1, 2020 | Time needed to complete other requirements is higher than expected | Degraded productivity, possible req failure | 4 | 3 | MEDIUM | Closed (3-21-2021) | - Do not implement reciprocal algorithm and use SALC to handle reciprocal case - Decided not to do SALC, reciprocal | Robot can still do prediction and avoidance | Technical Schedule | Technical Schedule | |||||||||||||||||
17 | 16 | Too many cooks in the kitchen | MP 6 | Tom, Calen | Dec 1, 2020 | Dec 10, 2020 | Avoidance algorithms don't work well together | - Loss in combined productivity - Trajectory Synthesis subsystem has issues combining and switching between trajectory avoidance algorithms | 4 | 3 | MEDIUM | Closed (3-21-2021) | - Demonstrate productivity increase from each algorithm separately (FVD) - Give trajectory synthesizer extra time in schedule (DONE) - Not doing trajectory synthesis (DONE) | - Can show each algorthm independly increases productivity | Programmatic Technical | Programmatic Technical | |||||||||||||||||
18 | 17 | CR-108 | Avoidance Takes Too Long | V4 | Tom, Yukun | Dec 1, 2020 | May 8, 2021 | Planning and avoidance algorithms take too long to process information | MP 6 is not satisfied | 1 | 1 | LOW | Open | - Reduce resolution of trajectory dictionary (Apr 1) - Optimize trajectory comparison and collision prediction algorithms - Create Alternative scaled down factory (Feb 1) - Implement memory management techniques (Apr 1) | Planning computation time is reduced | Programmatic Technical | Programmatic Technical | ||||||||||||||||
19 | 18 | Bad Ground Truth Data | R2 | Jinkun, Tom | Dec 8, 2020 | Sep 17, 2021 | Ground truth sensor system for RLS is not precise enough | Cannot validate positional accuracy of detection system in a dynamic environment | 5 | 2 | MEDIUM | Open | - Reserve money budget and time in schedule to upgrade camera. - Buy additional camera to use multiple cameras at once | - Can verify accuracy of detection system - Demonstration can be dynamic | Technical | Technical | |||||||||||||||||
20 | 19 | Stand Unstable | R2 | Jinkun, Tom | Dec 9, 2020 | Dec 9, 2020 | Ground truth subsystem camera may be unstable and fall over | Sensor for ground truth cannot be mounted safely, rendering ground truth system non-functional | 4 | 3 | MEDIUM | Open | - Have object which can weigh down the base - Create a protective casing for the camera | - Stand is stable - Camera is not destroyed if it accidently topples over | Technical Finance | Technical Finance | |||||||||||||||||
21 | 20 | AWS Bad Service | MP 6 | Yukun, Gitaek | Dec 9, 2020 | Mar 3, 2020 | Amazon web services is does not work well for our project for any reason | Cannot develop RLS system on a beefy compute platform | 5 | 3 | HIGH | Resolved | - Perform trade study on alternative platforms (Feb 1) - Gain basic understanding of how alternative platforms works (Feb 1) | We can switch to a different IaaS or PaaS platform quickly | Programmatic Technical Schedule Finance | Programmatic Technical Schedule Finance | |||||||||||||||||
22 | 21 | Bad Toy | Jinkun, Tom | Dec 9, 2020 | Dec 9, 2020 | A forklift or doll toy breaks | We cannot use the toy for obstacle detection | 3 | 2 | LOW | Resolved | Buy one additional toy of each kind | We can afford to loose a toy | Schedule Finance | Schedule Finance | ||||||||||||||||||
23 | 22 | Camera Too High | Jinkun, Tom | Dec 30, 2020 | Dec 30, 2020 | Camera on turtle bot is mounted too high | Tutrlebot has an unrealistic perspective of obstacles | 3 | 3 | MEDIUM | Resolved | Reconfigure the turtle bot to lower the camera | Technical | Technical | |||||||||||||||||||
24 | 23 | Assoc. Issues: | VRS/RLS Balance | Calen, Jinkun | Feb 4, 2021 | Mar 3, 2021 | There are a limited number of man hours that must be allocated between the VRS and RLS. Both system's may be greedy and cause a sub-optimal allocation | - VRS or RLS falls behind schedule - Spring Validation is reduced - VRS or RLS has too much work left to do in the fall | 4 | 3 | MEDIUM | Resolved (3-3-2021) | - Continuously evaluated progress in VRS and RLS. - Develop more precise Spring Validation plan - Prepare Jinkun to pivot from RLS to VRS when necessary (done) | VRS and RLS complete spring objectives If there are insufficient man hours, then VRS completes it's objectives and the RLS falls behind | Programmatic Schedule | Programmatic Schedule | |||||||||||||||||
25 | 24 | Unidirectional SALC | Tom, Calen | Feb 4, 2021 | Feb 4, 2021 | SALC causes non optimal behaviour because directional field makes robot "dodge" collisions by going in front of obstalce | Less efficient collision avoidance | 4 | 5 | HIGH | Closed (3-21-2021) | - Create modified SALC that is repulsive and does not lead the robot infront of obstacles | Programmatic Technical | Programmatic Technical | |||||||||||||||||||
26 | 25 | Rapid Avoidance Switching | Tom, Calen | Feb 8, 2021 | Feb 8, 2021 | Trajectory generator Jitters and switches between avoidance algorithms rapidy | Robot has erratic avoidance behaviour | 5 | 4 | HIGH | Closed (3-21-2021) | - Investigate Methods to reduce Jitter (Sept 1) - Potentially use cooldown time - Potentially use two threshold system | Technical | Technical | |||||||||||||||||||
27 | 26 | Gazebo Doesn't Like Magic | V4 | Yukun, Tom | Feb 8, 2021 | Mar 3, 2021 | Gazebo has issues when we directly set obstacle positions | Repeating obstacle trajectories is glitchy | 1 | 1 | LOW | Open | - Create reverse ackerman plugin for gazebo or use one off the shelf (Update 3/3/2021) | Forklift can repeat a trajectory successfully | Technical | Technical | |||||||||||||||||
28 | 27 | ROS Nav Integration | Yukun, Tom | Feb 12, 2021 | May 8, 2021 | Our custom avoidance algorithms do not integrate will with the ROS navigation stack | Cannot use our avoidance methods with ROS NAV stack | 5 | 3 | HIGH | Resolved (3-3-2021) | - Investigate ROS ASAP (Due 2-15-2021) - Flush out how intregation will work (DUE 2-17-2021) - Determine how to proceed (DUE 2-19-2021) | We either A) Integrate our avoidance algorithms with the ros Navigation stack or B) Create our own navigation stack | Programmatic Technical | Programmatic Technical | ||||||||||||||||||
29 | 28 | CR-340 | ROS Nav No Multi | Yukun, Tom | Feb 22, 2021 | Mar 3, 2021 | If we implement the global or local planner in Navigation stack, it might not be applicable to the multiple robot case | Cannot use Ros Global Planner | 5 | 1 | MEDIUM | Resolved (3-3-2021) | - As soon as the Global Planner is implemented, test it in a multi robot case - If a multi robot case is not ready, see if the the Global Planner can operate independently of each robot | We can determine if this issue materializes and fail fast | Use ETH/grid_map | Technical Schedule | Technical Schedule | ||||||||||||||||
30 | 29 | Gazebo Stability | V4 | Yukun, Tom | Mar 21, 2021 | May 4, 2021 | Gazebo has unstable obstacle motion | Simulation is unusable | 1 | 1 | LOW | Resolved | - Reduce obstacle weights - Reduce complexity of the simulation environment | Technical Schedule | Technical Schedule | ||||||||||||||||||
31 | 30 | Gazebo Doesn't Handle Collisions | V4 | Yukun, Tom | Mar 21, 2021 | May 4, 2021 | Gazebo simulation breaks or needs to be reset when a collision happens | Restoring gazebo to a stable state after a collision requires manual intervention and is takes time | 2 | 2 | LOW | Resolved | - Automatically restart scenario - Automatically abort and reboot simulation environment | - We know how to programmatically recover the simulation after a collision | Technical | Technical | |||||||||||||||||
32 | 31 | Gazebo is slow | V4 | Yukun, Tom | Mar 21, 2021 | May 4, 2021 | Gazebo simulation runs very slowly | Simulation | 2 | 3 | LOW | Resolved | - Get access to MRSD workstation - Test full scale environment on workstation - Create low poly cuboid version of factory | Simulation can run at a reasonable speed | Technical | Technical | |||||||||||||||||
33 | 32 | Occlusions | V4 | Calen, Tom | May 5, 2021 | May 5, 2021 | Occlusions prevent the robot from detcting obstacles early enough to avoid them | Robot has collisions which it predicted but could not avoid | 4 | 5 | HIGH | Open | - Make robot stop at intersections (PR 8) - Make forklifts slow down at intersections | Robot is more likely to observe an obstacle before it | Technical | Technical | |||||||||||||||||
34 | 33 | Multiple obstacles overloads gazebo | V4 | Calen, Tom | May 10, 2021 | May 11, 2021 | Gazebo is too slow when using multiple obstacles | Simulation can not be run on a locla machine | 4 | 4 | HIGH | Open | - Use paperspace web service to run simulation on the cloud | Programmatic Technical Finance | Programmatic Technical Finance | ||||||||||||||||||
35 | 34 | 3D Printed Parts Break | Jinkun, Tom | Sep 17, 2021 | Sep 17, 2021 | 3D Printed Parts break | Fan or camera cannot be mounted on turtlebot | 5 | 3 | HIGH | Open | - Print additional parts to have ready before anything breaks | |||||||||||||||||||||
36 | 35 | No Ciritical Path | Calen | Sep 19, 2021 | Sep 19, 2021 | Failure to do Gantt chart and depenency analysis will lead to project being incomplete at FVD | Multiple all nighters, flunking MRSD progam | 5 | 4 | HIGH | Closed (09-19-2021) | - Add all Sprints and Epics to Jira - Do Dependency Analysis on Roadmaps | -Improved foresight and project understanding leads to completing project on time without pulling all nighters | ||||||||||||||||||||
37 | 36 | No Battery Power | Jinkun | Oct 4, 2021 | Oct 4, 2021 | Jetson randomly powers off when on battery. Connection failure discovered. However, Jetson work ok when powered by tether. This means that there is a failure somewhere in the power system. | - Best case: Only connector needs to be replaced - Worst case: PDB is fried and needs to be replaced | 4 | 3 | MEDIUM | Open | - Sauder new connection and test ASAP - Buy new battery for OpenCR and run Jetson tethered | - We can determine if the issue is resolved by fixing the connection - RLS can run on Tether | ||||||||||||||||||||
38 | 37 | Test Environment Changed | Tom | Sep 25, 2021 | Oct 7, 2021 | Changes in the test environment on FVD cause RLS verification to fail | Expulsion from MRSD progarm | 5 | 3 | HIGH | Open | Add border to test environment so that obstacles have consistent backgrounds | No Expulsion | ||||||||||||||||||||
39 | 38 | CMU Wifi Drops | Yukun | Sep 28, 2021 | Oct 7, 2021 | CMU Secure wifi drops during FVD, causing RLS Verification to fail | Expulsion from MRSD progarm | 5 | 2 | MEDIUM | In Progress | Set up private Wifi network | No Expulsion | ||||||||||||||||||||
40 | 39 | RLS Camera Dies | Ykun | Sep 28, 2021 | Oct 7, 2021 | RLS camera dies during FVD, causing RLS verification to fail. Suspected causes are over heating or driver malfunction | Expulsion from MRSD progarm | 5 | 3 | HIGH | Open | - Debug ASAP - Develop design to install cooling fan for camera | No Expulsion | ||||||||||||||||||||
41 | |||||||||||||||||||||||||||||||||
42 | |||||||||||||||||||||||||||||||||
43 | |||||||||||||||||||||||||||||||||
44 |