S26 Micro Quests

Micro Quest Book - Summer 2026 (S26)

The Great Objective: Competition Ready

This term, Micro aims to complete the MVP to a competition-ready state while laying the groundwork for early stage learned sensor fusion. This will involve finishing our local planning stack, extensive real-world tuning, and building and integrating our Intel RealSense camera onto the car.

Term Objectives Summary

[Brief overview of what will be accomplished this term, organized into categories.]

  1. Hardware

    • Intel RealSense Camera Mount
  2. Software

    • SDF Based collision detection Node
    • Hybrid A* + path optmization Node
    • Intel RealSense Software integration
    • Race Ready Tuning
    • infrastructure improvements

Term Objectives and Scoring

Hardware

  1. Intel RealSense Camera Mount
ScoreCriteria
10/10Camera and mount can survive 8 consecutive laps with minor collision with No damage or minor scufs
7/10Mount is attached to platfrom + camera mounted on it
4/10Mount is Designed + Fabricated
0/10No mount is made

Minimum Requirements: There must be CAD files + a few printed prototypes for a score of 4/10

Software

  1. SDF Based collision detection Node
ScoreCriteria
10/10node can correct detect collision and suggest adjacent point to local planner
5/10node can correctly detect when a collision will happen
2/10Node can generate a correct SDF from a occupancy grid message
0/10Node does not exist in a functional state and produces no useful outputs

Minimum Requirements: Node must generate a correct SDF from occupancy grid for score of 2/10

  1. Hybrid A + Path Optimization Node*
ScoreCriteria
10/10Hybrid A* and path optimizer are combined and working reliably on the car
7/10Hybrid A* and path optimizer are combined and validated in simulation
4/10Path optimizer is built out and producing optimized paths independently
0/10Hybrid A* does not output a valid path

Minimum Requirements: Hybrid A* must output a valid path for a score of 4/10.

  1. Intel RealSense Software Integration
ScoreCriteria
10/10ROS driver is installed, fully functional, and capable of automatically recording rosbags with depth, point cloud, and image data
7/10Camera produces depth maps, point clouds, and RGB images through the ROS driver
4/10Camera outputs raw image and depth data successfully
0/10No usable output from the camera

Minimum Requirements: Camera must produce raw image and depth output for a score of 4/10.

  1. Race-Ready Tuning
ScoreCriteria
10/10Car is reliably running at a top speed of 10 m/s
7/10Car is reliably running at a top speed of 8 m/s
3/10Car is reliably running at a top speed of 5 m/s
0/10Car is stuck at 3 m/s or below

Minimum Requirements: Car must reach a top speed of 5 m/s for a score of 3/10.

  1. Infrastructure Improvements
ScoreCriteria
5/5Foxglove is working correctly and well-documented, and an automated rosbag recording script is built that saves bags with timestamped run names to a designated location on launch
2/5Foxglove is set up, working correctly, and well-documented
0/5No functional infrastructure improvements delivered

Minimum Requirements: Foxglove must be working correctly and documented for a score of 2/5.


Scoring Template

Hardware

Quest NameDescriptionScore
Intel RealSense Camera MountIntel RealSense Camera mounted to car and ready to use in race

Software

Quest NameDescriptionScore
SDF Based Collision Detection NodeBuild a node that generates SDFs from occupancy grids and detects collisions to assist the local planner
Hybrid A* + Path Optimization NodeDevelop and integrate a Hybrid A* path planner with a path optimizer, validated in sim and on the car
Intel RealSense Software IntegrationInstall ROS drivers and integrate the Intel RealSense camera to produce depth, point cloud, and image data
Race-Ready TuningTune the full stack for competition-level performance, targeting up to 10 m/s top speed
Infrastructure ImprovementsSet up Foxglove with documentation and build an automated rosbag recording script on launch