Week 56: collision avoidance: it’s done! (Patreon)

his week I finished the relativistic implementation of the new collision avoidance algorithm. The new code is much better than the old one. Well, this is what I expected after working on the classical proof of concept a couple of weeks back. Here is the commit message:

Author: Matteo Franchin <matteo.franchin@relativisticgame.net>

AuthorDate: Mon Nov 20 18:34:06 2023 +0000

Commit: Matteo Franchin <matteo.franchin@relativisticgame.net>

CommitDate: Wed Nov 22 09:51:56 2023 +0000

engine: new collision avoidance algorithm

Implement a more refined version of the collision avoidance algorithm.

The new algorithm has the following advantages over the older one:

- objects are not bouncing back when hitting walls. Objects should now

   accurately slow down close to walls.

- they can now go faster, especially closer to walls. Objects can

   move fast and parallel to walls.

- a maximum velocity can be set for objects, independently on whether

   they are colliding with walls

- the algorithm is much better at preventing objects from traversing

   walls. It takes care of time-discretisation to achieve that.

- friction is introduced to spare the player from continuously braking

   and correcting the direction of travel.

I attach a video that tries to show some of these claims (note how the robot goes fast parallel to walls, without bouncing), although a video isn’t the best way of assessing these sorts of changes. An intermediate game demo (say 1.1) would work better. Ideally, it’d be nice to allow people to try different parameters. For example, I can make the robot go faster, but then navigation becomes less smooth inside narrow corridors. It is a bit of a trade-off. In the end, the parameters should be set in a way that improve user-experience. Here is where feedback could help. The problem is that my resources are limited. I have to carefully decide where to invest my time.

On that note, while thinking about levels for the game I came out with a super cool paradox on time-dilation. I really want to make a YouTube video out of it, but - at the same time - I don’t want to delay demo 2.0 too much.

Also, this week I finally could go back to working on game mechanics. I am introducing new types of keys and keyholes. I hope to make significant progress in the next weeks and design room 2 and room 3. I also plan to work on save/load functionality (as it could be handy for development.) After that, the next thing in my TO-DO list is working on sound-effects.

Conclusion: implementing the collision avoidance algorithm took much longer than I expected. Not great. There is a positive side to it, though. The new algorithm solves one problem that I noticed previously and would have taken time to fix. The door was dragging the player when moving. I expected the new algorithm to behave better, but I wasn’t sure that the problem would completely go away. It seems that is the case. Hooray!