POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models package
Submodules
POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_agent_models module
Tests for the agent-channel observation model (agent_models.py).
- POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_agent_models.test_log_probability_penalizes_seeing_an_occluded_agent()[source]
Reporting an occluded agent is scored as near-impossible.
Purpose: Validates occlusion gating in the agent-channel density
Given: A target occluded by a blocker on the ego->target sight line When: log_probability scores an observation that (wrongly) reports the occluded target
vs one that correctly omits it
Then: The correctly-omitted observation scores far higher
Test type: unit
- POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_agent_models.test_perceive_gates_out_of_range_agent()[source]
perceive keeps a near agent and zeroes an out-of-range one.
Purpose: Validates the range gate on the agents channel
Given: A model with a 50 m range and a near + far agent (both present) When: perceive is called on the flat agents block Then: The near slot stays present and the far slot is zeroed
Test type: unit
- POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_agent_models.test_render_noiseless_preserves_pose()[source]
render(noisy=False) gates without perturbing a detected pose.
Purpose: Validates the noise-free rendering path used to build clean observations
Given: A model with no range gate and a single detected agent with a nonzero pose When: render is called with noisy=False Then: The agent’s pose is returned unchanged
Test type: unit
POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_gnss_models module
Tests for the GNSS-channel observation model (gnss_models.py).
- POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_gnss_models.test_log_probability_prefers_the_truth()[source]
The density scores the true reading above a shifted one.
Purpose: Validates the GNSS Gaussian density ranks the truthful reading higher
Given: A GnssObservationModel and a clean 2-D reading When: log_probability scores the truth vs a 3 m-shifted reading Then: The truthful reading has the higher log-probability
Test type: unit
- POMDPPlanners.tests.test_environments.test_carla_pomdp.test_carla_perception.test_observation_models.test_gnss_models.test_perceive_returns_two_vector()[source]
perceive returns a 2-D GNSS reading.
Purpose: Validates the GNSS sampler shape and channel handling
Given: A GnssObservationModel and a clean 2-D reading When: perceive is called Then: A length-2 array is returned
Test type: unit