watonomous.github.io

[ Software Division : W20.0 Path Planning EOC ]

Created by [ Charles Zhang] on Feb 12, 2020

Link to EOC guide

Heading

Name: Charles Zhang

Date: January 29, 2020

Development Cycle: 0

Division: Software

Group: Path Planning

Status

[GREEN]

Seem to be on track for milestones.

OKR Timeline Progression

[W20.O1] Trajectory Planning

[W20.O1.KR1] Frenet Obstacle Avoidance

0% -> 10% Complete

[W20.O1.KR3] Spline creation

0% -> 10% Complete

Ben, Anqi, and Tyson are reading through sample code provided by Tyson (CppRobotics) implementing spline creation and frenet planning.
Ambitious milestone for a prototype is End of dev cycle 1

[W20.O2] Routing

[W20.2.KR1] Modify routing algorithm to tag the intersections

[W20.O1.KR2] Communicate waypoints (intersections) with behavioural planner

[W20.O1.KR3] Update Global Map using HD Map roundabouts and signs

[W20.O1.KR5] Create SD maps

0% → 0% complete

John, David, and Lilly are still onboarding on the global planner code.
Milestone for a working prototype of global planner: End of dev cycle 2
Intersection behaviour is blocked by processing HD map parsing and filtering

[W20.O1.KR4] Stop and shift into park for 5s once you’ve reached a destination

0% → 10% complete

Talked with CAN interfacing about gear change requirements.

[W20.O3] Feedback Controllers

[W20.O3.KR1] Add obstacle avoidance to MPC code

0% → 0% complete. Not currently a priority

[W20.O3.KR2] Investigate and improve braking

0% → 20% complete

Connor is working on this. Tried some experimental fixes at the test track
Milestone for this OKR is 2 dev cycles.
Currently blocked by lack of knowledge of how we should solve the problem. Will talk to Mohammed Daoud, Tyson, and other Mohamed
Milestone for proper braking: End of dev cycle 1

[W20.O4] Simulation Challenge

0% → 10% complete

Charles is working on connecting a MATLAB/SIMULINK vehicle model to our ros nodes. We can use this for closed-loop testing using RVIZ as a visualizer
Need to investigate resources that MATLAB provided in the simulation workshop
Using a simulink vehicle model seems completely feasible, and will probably encompass the extent of our simulation challenge efforts. Stretch goal would be to write a parser for the scenario builder, but currently seems like more work than it's worth
Milestone for vehicle model completion: End of dev cycle 1

[W20.O5] Potential Fields

0% → 0% complete

On pause as we think that the frenet planner will be a better choice. Will work on potential fields once MPC braking is fixed

[W20.O6] Behavioural Planner

[W20.O6.KR1] Integrate lane changing with Global Planner

Blocked by global planner
Milestone: Dev Cycle 2

[W20.O6.KR2] Improve Lane Changing

0% → 20% complete

Talked with CAN about blinkers
Ray is working on blinker logic
Spencer has begun work on lane-changing obstacle avoidance
Milestone for prototype: End of Dev Cycle 1
Milestone for completion: End of Dev Cycle 2

[W20.O6.KR3] Implement lane logic based on lane type

[W20.O6.KR4] Verify Dynamic Obstacle behaviour

All blocked by Perception/processing no traffic lights or processing HD map filtering.
No members currently assigned to this task (TBD)

[W20.O7] Infrastructure

[W20.O7.KR1] Create Unit Tests for ros nodes

0% → 80% complete

Unit tests for Trajectory Planner, Behavioural Planner, Occupiable objects are done.
John is still working on getting unit tests for Global planner.
Deadline: End of dev cycle 1

[W20.O7.KR2] Create Integration tests to replace autonomous mode pre-checks

[W20.O7.KR1] RViz Improvements

These are to be handed off to Infra

Link To Retrospective

Retro

Document generated by Confluence on Dec 10, 2021 04:01

Atlassian