Automate Testing for UAV Obstacle Avoidance

hi my name is jin jin soon i'm a developer for uav toolbox in this demo i'm going to show how requirements are linked to your uav applications in simulink and how to verify these requirements assuming that you have previously experimented with an algorithm that guides uav to avoid obstacle while trying to transport package from point a to point b now you want to productize this feature as the very first step of productization you can alter requirements for your delivery system using simulink requirement editor and link it with your simulink model implementations here are some existing requirements on the uavs guidance system and we can add new requirements for the obstacle avoidance feature in this new requirement we require the uav to always keep a safe distance to other obstacles in the environment which is a very broad requirement that we are going to use as demonstration in this video we implemented the obstacle avoidance algorithm using vector field histogram block from navigation toolbox based on lidar readings from unreal code simulation and we can link our requirements to this implementation once we have our requirements defined we can add automated tests to verify such requirements using simulink tests we can extract the uavs algorithm and simulation section into a test hardness and add our tests i have already added one test case ahead of time to lock down that the uav's minimum distance two obstacles detected by lidar must remain higher than 0.5 during the simulation this test verifies our requirement that uav must always be away from obstacles with one click we can execute our test in simulation and verify the algorithm behavior the test result is all green meaning that our uav can keep its distance and in the requirement table we now see that the safety distance requirement is implemented and verified shown as blue and green this helps us keep track of complicated feature development and verification we can add more tests that check algorithms behavior with different goal locations in the virtual city environment and verify that the safety distance is always observed by the uav additional requirements can also be added to verify the uav's takeoff and landing behavior as well for example uav should be able to check that there are no surrounding obstacles before taking off and we can create new tests that place even near these obstacles to validate the takeoff behavior as this video demonstrated using uav toolbox simulink requirements and simulink tests products enable quick iteration over algorithm requirement definition implementation simulation and verification we hope these features can help in your future projects to develop efficient and safe algorithms for uav navigation thank you

Channel: MATLAB

Views: 1,247

Rating: 5 out of 5

Keywords: MATLAB, Simulink, MathWorks

Id: lvYRSjcZqoA

Length: 3min 42sec (222 seconds)

Published: Wed Sep 22 2021