mpccbf_test.py solve failed

[luozhexin]

System information
OS Platform and Distribution (e.g., Linux Ubuntu 18.04): Linux Ubuntu 20.04
Branch: master
python3.9(no python3.6 due to cvxopt install failed)
Steps to reproduce the issue:
python car_racing/test/mpccbf.py --track-layout l_shape --simulation --plotting --animation

Error detail:
Car-Racing/src/policy.py", line 224, in calc_input
sol = opti.solve()
File "/home/wsco/anaconda3/envs/car_racing/lib/python3.9/site-packages/casadi/casadi.py", line 47783, in solve
return _casadi.Opti_solve(self, *args)
RuntimeError: Error in Opti::solve [OptiNode] at .../casadi/core/optistack.cpp:157:
.../casadi/core/optistack_internal.cpp:998: Assertion "return_success(accept_limit)" failed:
Solver failed. You may use opti.debug.value to investigate the latest values of variables. return_status is 'Infeasible_Problem_Detected'

[hesuieins]

Did this error occurs consistently? Since random noise is included in the simulator, the optimization problem could be infeasible in some cases. Can you try with the option zero_noise to see if the error occurs again?

[luozhexin]

Thanks for your reply, but the problem is still not solved.
I made the following modifications to the code:
1、In mpccbf_test.py add
ego.set_zero_noise()
car1.set_zero_noise()
car2.set_zero_noise()
in if args["simulation]
2、In base.py
self.zero_noise_flag = True

It will still report the same error as before，Only by making the following modifications can the program run
simulator.sim(sim_time=5.0)
If the simulation time is set to 10, an error may still be reported.

[junzengx14]

One general suggestion, could we specify a few ubuntu versions in our CI setup?

[junzengx14]

@hesuieins @yifanzeng0408 Are you tracking this issue?

[Bubblefn]

@luozhexin Have you solved this problem? I also encountered the same problem.

[yifanzeng0408]

@luozhexin Have you solved this problem? I also encountered the same problem.

I have run the test and didn't encounter the problem. My guess is that the solver fails because of conflicting obstacle avoidance constraints. Maybe you can try deleting car1 and see the results?

[yifanzeng0408]

@junzengx14 the fixes of this issue is proposed in this PR.

[junzengx14]

@Bubblefn @luozhexin Are the problem resolved with above patches?

[luozhexin]

@junzengx14 Yes, it solved the problem, thank you