Sparkbot - Maze Solving Micromouse

Tools & methods

• Flood-fill path planning
• Embedded programming
• Robot assembly

Project overview

We built Sparkbot as Team SparkBots’ maze-solving robot for Robofest 2023, organized by SLIIT. I implemented the flood-fill navigation algorithm and worked on constructing the robot with the team.

The project brought physical robot assembly and embedded navigation together. The robot needed to move through the maze while maintaining a useful representation of the available routes.

My contribution & team

My contribution included implementing flood fill and building the robot. This connected the route-planning logic to a physical platform, where the decisions made by the program had to correspond to the robot’s position and movement.

The team included Janitha Karunarathna, Chamika Dilshan, Sasini Perera, and Hansika Karunathilake alongside my participation. The supplied robot photograph records the assembled platform, and the maze photograph records the competition environment.

How flood fill supports maze solving

Flood fill represents the maze as connected cells and assigns distance values outward from the goal. In a maze with equal-cost moves, these values describe the number of cell steps to the goal through the currently known connections.

1. Represent cells, their connections, and observed walls in the maze map.
2. Set the goal distance to zero and propagate distances through reachable neighbouring cells.
3. Choose a reachable neighbour with a lower distance value.
4. Move toward that cell and update the map as new walls are observed.
5. Recompute distances when new information changes the available routes.

This explains the navigation principle used in the project. The robot’s path choices depend on the maze information available at each step, so exploration and route planning work together.

Robot construction & system integration

The supplied photograph shows a compact robot with a layered chassis, interconnected electronics, sensor modules, and drive wheels. Building the robot brought these components together into one mobile platform.

For maze navigation, physical movement and the map must stay consistent. Cell-to-cell travel, turns, wall observations, and software decisions need to be coordinated so the next route choice corresponds to where the robot actually is.

The project therefore combined programming with practical assembly and integration, rather than treating the navigation algorithm as a standalone exercise.

Competition environment

The competition maze uses a grid of passages separated by walls. This makes wall layout, junctions, and reachable neighbouring cells the key information for navigation.

The maze provides the physical setting for the flood-fill implementation: each observed passage changes the set of routes available to the robot.

Final-round qualification

The supplied certificate of achievement names W. S. R. K. Perera of Team SparkBots and confirms qualification for the final round of Robofest 2023 in the university category.

The event was organized by the Department of Electrical and Electronic Engineering at the Sri Lanka Institute of Information Technology, Malabe. The certificate records finalist qualification.

Project outcome

The project produced a built robot, a flood-fill navigation implementation, and a university-category finalist result. It gave me practical experience in connecting a maze-solving algorithm to a physical robot and working with a team in a competition setting.

Flood-fill navigation loop

An overview of the documented project scope.

1. 01Observe maze
2. 02Update map
3. 03Flood-fill distances
4. 04Select & move

Original robot photograph

Project resources

The supplied robot and maze photographs, project notes, and certificate support the project overview and finalist result.