In order to get a faster speed compared to the baseline, the first test was intuitively to increase the speed in CC.cpp. This, however, resulted in the car steering in a very unstable way and sidetracking off the road. To fix this, the execution time and period should be reduced to check and adjust the car’s position with a higher frequency.
Baseline: The baseline model uses a period time of 500 ms, utilization 10% for LK and CC, and utilization of 20% for the remaining four jobs. This gives a period of 500 ms with total utilization of 100% (210 + 420). The baseline model's track time was recorded to 2:24.
To improve this, it is desirable to shorten the period and execution times. This is especially true for steering (LK), in order to maintain a smooth and stable course through the track. Another function to keep at a high update frequency is CC to make sure that the speed is as constant as possible.
Improvements: The first parameter to change was the period and deadline times. Given the restraints in the project, it is obvious that at maximum one job can have execution times <= 100ms (since it gives utilization of 50%). For somewhat simpler calculations, the LK and CC were decided to have the same execution and period times at half of the remaining tasks. Since a total period time of 200 ms would lead to the LK and CC having a 100 ms period (not possible as described above), 300 ms was chosen instead. This gives LK and CC a period time of 150 ms. Using the lowest allowed utilization for all jobs resulted in all jobs having 30 ms execution times. Total utilization is 80% (220 + 410). This allowed for the speed in CC.cpp to be increased to 70 kph (75% increase). After running the simulation, the best lap time was recorded to 1:28, which is almost a minute improvement.