[Autoware Labs] Dense Urban ODD - Reviewing and Tracking Failed Scenarios

[brkay54]

Checklist

• I've read the contribution guidelines.
• I've searched other issues and no duplicate issues were found.
• I've agreed with the maintainers that I can plan this task.

Description

Follow-up task of:

• Create Dense Urban ODD Scenarios autoware#4636

Failed Scenario Summary:

• Development Tracking Sheet

Objective

This issue tracks the progress and improvements related to the scenario catalog created for autonomous driving tests within a Dense Urban Operational Design Domain (ODD). The primary objectives are:

• Defining Problems: Clearly define the problems for developers using the scenarios below.
• Tracking Development: Monitor the development process.

Dense Urban ODD Characteristics

• Low-Speed Driving: Typical speeds around 15 km/h.
• Complex Traffic Elements: Includes dense traffic with various dynamic interactions.
• Dynamic Agents: Moving pedestrians and cyclists are frequently encountered.
• Bidirectional Movement: Vehicles operate in lanes supporting bidirectional flow.
• Traffic Rule Non-compliance: Other road users may exhibit behaviors such as:
  • Driving in the wrong direction.
  • Cutting in unexpectedly from off-road areas.
  • Pedestrians jaywalking.
  • Drivers overtaking improperly.

Task for Developers

• Investigating failed scenarios: Identify root causes and potential fixes for failed scenarios.
• Improving Software Packages: Enhance functionalities based on scenario feedback.
  • Fine-Tuning Parameters: If the package already has the functionality to handle the failed case, fine-tune the package parameters.
  • Proposing Improvements: If the package has the functionality but it is insufficient to handle the failed case, propose improvements to the functionality.
  • Adding New Functionalities: If the package lacks the functionality to handle the failed case, propose adding new functionality to the target package.
• Creating Additional Scenarios: Develop new scenarios to test various other conditions and ensure robustness.

List of the scenarios

Click to expand the list of the scenarios

Useful Links

• Scenario Suit: Autoware Evaluator
• Scenario Description Document: Google Presentation

- [Development Tracking Sheet]( https://docs.google.com/spreadsheets/d/1WTqiyUndtdBNgiNnAApjOXhma-QAU_SBPdD626pMr2Q/edit?gid=0#gid=0 )

UC-NTR-001-0001

• Scenario Description: Ego vehicle is overtaken from outside of the road by a faster moving NPC.
• Development Focus: Ensure ego vehicle remains stationary until the high velocity NPC has completed its overtaking maneuver.
• Scenario Links:
  • UC-NTR-001-0001_case1
  • UC-NTR-001-0001_case2
• Map Link: LEO-VM-0001
• Issue/PR Links:

UC-NTR-001-0002

• Scenario Description: Ego vehicle is overtaken from outside of the road by a slower moving NPC.
• Development Focus: Ensure ego vehicle departs before the NPC completes its overtaking maneuver because its velocity is too low.
• Scenario Links:
  • UC-NTR-001-0002_case1
  • UC-NTR-001-0002_case2
• Map Link: LEO-VM-0001
• Issue/PR Links:

UC-NTR-001-0003

• Scenario Description: Ego vehicle is overtaken from the same direction lane by a faster moving NPC before starting.
• Development Focus: Ensure ego vehicle remains stationary until the high velocity NPC has completed its overtaking maneuver.
• Scenario Links:
  • UC-NTR-001-0003_case1
  • UC-NTR-001-0003_case2
• Map Link: LEO-VM-00009
• Issue/PR Links:

UC-NTR-001-0004

• Scenario Description: Ego vehicle is overtaken from the same direction lane by a slower moving NPC before starting.
• Development Focus: Ensure ego vehicle departs before the NPC completes its overtaking maneuver because its velocity is too low.
• Scenario Link:
  • UC-NTR-001-0004_case1
  • UC-NTR-001-0004_case2
• Map Link: LEO-VM-00009
• Issue/PR Links:

UC-NTR-001-0005

• Scenario Description: Ego vehicle is overtaken from the opposite direction lane by a faster moving NPC before moving.
• Development Focus: Ensure ego vehicle allows the high velocity NPC to overtake.
• Scenario Link: UC-NTR-001-0005
• Map Link: LEO-VM-00010
• Issue/PR Links:

UC-NTR-001-0006

• Scenario Description: Ego vehicle is overtaken from the opposite direction lane by a slower moving NPC before moving.
• Development Focus: Ensure ego vehicle departs before the NPC completes its overtaking maneuver because its velocity is too low..
• Scenario Link: UC-NTR-001-0006
• Map Link: LEO-VM-00010
• Issue/PR Links:

UC-NTR-001-0007

• Scenario Description: Ego vehicle is overtaken from the same direction lane by a faster moving NPC while driving.
• Development Focus: Ensure ego vehicle maintains safe distance and trajectory during overtaking.
• Scenario Link: UC-NTR-001-0007
• Map Link: LEO-VM-00009
• Issue/PR Links:

UC-NTR-001-0008

• Scenario Description: Ego vehicle is overtaken from the same direction lane by a faster moving NPC which then slows down.
• Development Focus: Ensure ego vehicle adapts to changes in NPC's speed without significant deceleration.
• Scenario Link: UC-NTR-001-0008
• Map Link: LEO-VM-00009
• Issue/PR Links:

UC-NTR-002-0001

• Scenario Description: Ego vehicle overtakes a stopping or slow-moving NPC on the side of the road.
• Development Focus: Ensure safe overtaking of stationary or slow-moving obstacles.
• Scenario Links:
  • UC-NTR-002-0001_case1
  • UC-NTR-002-0001_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-NTR-002-0002

• Scenario Description: Ego vehicle overtakes NPCs parked on different sides of the road.
• Development Focus: Navigate and maneuver around multiple parked obstacles safely.
• Scenario Links:
  • UC-NTR-002-0002_case1
  • UC-NTR-002-0002_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-NTR-002-0003

• Scenario Description: Ego vehicle overtakes moving NPCs on the side of the road.
• Development Focus: Ensure dynamic decision-making for overtaking moving obstacles.
• Scenario Links:
  • UC-NTR-002-0003_case1
  • UC-NTR-002-0003_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-NTR-002-0004

• Scenario Description: Npc is overtaking another Npc in opposite side. It suddenly appears on Ego's trajectory.
• Development Focus: Ensure Ego can slowdown for vehicle comes from opposite direction.
• Scenario Link: UC-NTR-002-0004
• Map Link: LEO-VM-0010
• Issue/PR Links:

UC-NTR-003-0001

• Scenario Description: Ego vehicle decides to stop or pass in an intersection while NPC interferes with its trajectory.
• Development Focus: Handle intersection if intersection module is not triggered.
• Scenario Link: UC-NTR-003-0001
• Map Link: Nishi-Shinjuku
• Issue/PR Links:

UC-NTR-003-0002

• Scenario Description: Ego vehicle decides to stop or pass in an intersection while NPC interferes with its trajectory during a turn.
• Development Focus: Ensure safe intersection handling during turning maneuvers. Ensure there won't be any unnecessary stopping.
• Scenario Link: UC-NTR-003-0002
• Map Link: Nishi-Shinjuku
• Issue/PR Links:

UC-NTR-003-0003

• Scenario Description: Ego vehicle decides to stop or pass in an intersection while NPC interferes with its trajectory from the right.
• Development Focus: Ensure intersection module is working for low velocity road users.
• Scenario Link: UC-NTR-003-0003
• Map Link: awf_cicd_virtual_E_dev
• Issue/PR Links:

UC-NTR-004-0001

• Scenario Description: Ego vehicle navigates a bidirectional lane with no line markings.
• Development Focus: Ensure safe navigation in undivided lanes with oncoming traffic.
• Scenario Link: UC-NTR-004-0001
• Map Link: LEO-VM-00038
• Issue/PR Links:

UC-NTR-005-0001

• Scenario Description: Ego vehicle cruises behind a front vehicle.
• Development Focus: Maintain safe following distance and speed adjustment.
• Scenario Link: UC-NTR-005-0001
• Map Link: LEO-VM-00001
• Issue/PR Links:

UC-VRU-001-0001

• Scenario Description: Ego vehicle passes over a group of pedestrians.
• Development Focus: Navigate pedestrian dense areas without collisions.
• Scenario Links:
  • UC-VRU-001-0001_case1
  • UC-VRU-001-0001_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-VRU-001-0002

• Scenario Description: Ego vehicle slows down for an unexpected pedestrian.
• Development Focus: React to sudden pedestrian crossings.
• Scenario Link: UC-VRU-001-0002
• Map Link: LEO-VM-00010
• Issue/PR Links:

UC-VRU-002-0001

• Scenario Description: Ego vehicle allows a bicycle to drive first.
• Development Focus: Ensure safe interaction with bicycles on the road.
• Scenario Links:
  • UC-VRU-002-0001_case1
  • UC-VRU-002-0001_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-VRU-002-0002

• Scenario Description: Ego vehicle drives through destination while a bicycle is present.
• Development Focus: Manage overtaking and following scenarios involving bicycles.
• Scenario Links:
  • UC-VRU-002-0002_case1
  • UC-VRU-002-0002_case2
• Map Link: LEO-VM-0001
• Issue/PR Links:

UC-VRU-002-0003

• Scenario Description: Ego vehicle drives through destination while a bicycle is driving on the side of the road.
• Development Focus: Ensure smooth interaction with driving bicycles while avoiding it.
• Scenario Links:
  • UC-VRU-002-0003_case1
  • UC-VRU-002-0003_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-VRU-002-0004

• Scenario Description: Ego vehicle drives through destination while a bicycle is approaching from outside the road.
• Development Focus: Decelerate comfortably to avoid collisions.
• Scenario Link:
  • UC-VRU-002-0004_case1
  • UC-VRU-002-0004_case2
• Map Link: LEO-VM-0008
• Issue/PR Links:

UC-VRU-002-0005

• Scenario Description: Ego vehicle drives through the destination while two bicycles drive in the same direction.
• Development Focus: Wait for NPCs to pass before overtaking.
• Scenario Link: UC-VRU-002-0005
• Map Link: LEO-VM-00038
• Issue/PR Links:

Purpose

The primary goal of this task is to enhance the capability of autonomous driving software for low-speed driving in dense urban areas with mixed road users. This involves improving the system's awareness of various conditions and its responsiveness to these conditions. By fine-tuning trajectory planning, cruise control for dynamic obstacles, adaptive cruise planning, and avoidance strategies, we aim to ensure smoother, safer, and more efficient driving in complex urban environments.

Possible approaches

Improvement Target	Description	Related Scenarios	Assignees	Development
Path Planning and Trajectory Adjustment	- Adjust the trajectory to align more closely with the right or left edge of the lane for bi-directional lanes, depending on the direction of traffic flow.	- UC-VRU-002-0005 - UC-NTR-004-0001	- @beyzanurkaya	Issue #7923
Cruise Planning for Dynamic Obstacles	- Expand the coverage area to consider objects further away from the trajectory. - Autoware should react much earlier if the predicted paths of other vehicles intersect with our planned path to avoid potential collisions. - Avoid unnecessary reactions to objects that briefly enter and exit our path to prevent abrupt and unnecessary slowdowns, ensuring smoother and more efficient driving.	- UC-NTR-001-0001 - UC-NTR-001-0002 - UC-NTR-001-0003 - UC-NTR-001-0004 - UC-NTR-001-0005 - UC-NTR-001-0006 - UC-NTR-002-0004 - UC-NTR-003-0001 - UC-NTR-003-0002 - UC-NTR-003-0003 - UC-VRU-001-0002 - UC-VRU-002-0001 - UC-VRU-002-0004	- @brkay54	Issue #7864
Adaptive Cruise Planning for Following Obstacles	- The ego vehicle may overshoot its velocity while following a dynamic object; find a better approach for velocity planning while following an object. - Maintain a following distance within 2-3 seconds.	- UC-NTR-001-0007 - UC-NTR-001-0008 - UC-NTR-005-0001 - UC-VRU-002-0002	- @ahmeddesokyebrahim	Issue #7877 Issue #7922 Issue #8311
Static and Dynamic Avoidance	- The system needs better capabilities to dynamically avoid objects coming from the opposite direction. Test and improve dynamic avoidance to avoid oncoming objects based on their future positions. - Test current avoidance and develop enhanced safe avoidance capabilities for vulnerable road users.	- UC-NTR-002-0001 - UC-NTR-002-0002 - UC-NTR-002-0003 - UC-NTR-002-0004 - UC-NTR-004-0001 - UC-VRU-001-0001 - UC-VRU-002-0003	- @ahmeddesokyebrahim - @beyzanurkaya	Issue #7634 PR #1024 Issue #8539
Intersection with low speed NPC	- The system must ensure collision avoidance with oncoming NPCs at low speeds within intersection areas when NPCs intersect the Ego vehicle's path.	- UC-VRU-003-0003	- @ahmeddesokyebrahim	Issue #7080

NOTE: A scenario might be in multiple improvement targets because the scenario needs both improvements to pass the evaluation.

Definition of done

• Problem Identification and Solution Proposal: Developers have created issues that clearly define the problem and propose a viable solution.
• Implementation of Improvements: All necessary improvements have been made to the relevant software packages.
• Addition of New Scenarios: Additional scenarios have been added into Autoware Evaluator to test other various conditions.
• Successful Scenario Execution: All scenarios have been executed successfully, with the system passing each one without issues.

[mitsudome-r]

Failed Scenario Summary: https://docs.google.com/spreadsheets/d/1WTqiyUndtdBNgiNnAApjOXhma-QAU_SBPdD626pMr2Q/edit?gid=0#gid=0

Used Map: https://docs.google.com/spreadsheets/d/1Uvgel4X-1vBUw-auISpRZ5-TUkvWJVV_CrL7MYe6HPg/edit?gid=0#gid=0