05Rover mission-planning softwareActive
MPT
Mission Planning Tool
- 3person team
- Anylaptop, even with low specs
Part of my CSA internship: while the rest of the team worked on the legacy software, I built a new rover mission-planning tool that runs on any laptop, low-spec included, and explored unconventional ways to implement path planning.
- Role
- Front end & path-planning R&D
- Team
- 3-person team
- Period
- Sep 2026 — Present
- With
- Canadian Space Agency · Internship
Illustrative demo · MPT
Loading instrument…
- 3
- person team
- Any
- laptop it runs on, even with low specs
- CSA
- internship project
01The brief
A planning tool for any laptop
As part of my internship at the Canadian Space Agency, I worked on rover mission planning: software that helps decide where a rover should go, and how it should get there.
We were a team of three. The other two worked on the existing, legacy software. I built a new tool alongside it, with two aims: that it should run on any laptop, low-spec ones included, and that it should try some unconventional ways of implementing path planning.
Because the work was done at the CSA, this page describes it in general terms. The interactive planner at the top is not that tool: it is a small illustration I wrote for this site, on made-up terrain, to show the kind of problem path planning solves.
- Context
- Canadian Space Agency · internship
- Team
- 3 people
- My part
- Front end · path-planning R&D
- Aim
- Runs on any laptop, even a low-spec one
02My role
The front end, and a different way to plan
- 01
Own the front end
Designed and built the interface of the new tool: what a planner sees and touches when setting up a plan and checking it.
- 02
Look beyond the usual
Explored unconventional ways of implementing path planning, rather than only the conventional ones.
- 03
Keep it light
Built the tool to run on any laptop, including low-spec machines.
- 04
Work in a team of three
Two teammates worked on the legacy software; I worked on a new tool.
03How planning works
From terrain to route, in general terms
Most grid-based path planning comes down to three ideas. The planner at the top of this page uses all of them, on terrain that does not exist.
01
Terrain becomes a grid
Planning starts by turning the ground into cells and asking, for each one, how hard it is to cross: how steep it is, how close it sits to a hazard, how much light it gets.
02
Costs decide the route
A search then finds the cheapest chain of moves between two points. “Cheapest” is a choice: least time, safest ground, least energy. The same search with different costs gives a different route.
03
Time changes the answer
Light moves, so a route that works now can be in shadow ten minutes later. A planner can take that into account, and even choose to wait.
04About the demo
A small illustration
It is not the CSA’s software. The terrain is made up, its parameters are illustrative, and it uses no CSA data or code.
It is simplified on purpose: the terrain is procedural, the patch is about a kilometre across, the sun moves far faster than the real one so that a short drive shows what takes hours, and there is no battery model.
Things to try
- Move the pins
- Drag A and B, or focus one and use the arrow keys.
- Fastest or Safest
- Same terrain, two definitions of “best”: watch the route change.
- Draw a keep-out zone
- Choose Keep-out zone, then click, tap or drag on the map to close ground off. The route goes around it.
- Sunlight only
- The rover plans around the light, and waits where it has to.
- Search method
- Under Advanced: A* and Dijkstra return the same route. Turn on Explored to see how much less map A* reads.