E11 – Find a lost dog with a drone

General description. The aerial robot must perform a surveillance flight looking for a dog that a customer has reported it was lost in the mall centre. The robot must be able to fly autonomously around the area, detect the lost dog and report its location. There can be obstacles in the flying area that must be avoided during the flight. Moreover, the robot is part of a “monitoring” infrastructure that should report about the status of the mall continuously while performing other duties. Two other objects may be present on the ground, such as a garbage bag and a cardboard box. The robot should report their position if one or more of these are spot while looking for the dog. Of course, the robot should report and update continuously the position of the dog

Platforms allowed. VTOL aerial robot that can navigate in outdoor and indoor spaces, detect and avoid obstacles.

Setting. The episode must take place inside the flying arena. The lost dog can be represented by a teddy dog mounted on a small RC car, so it can be moved around the arena.

Procedure. The request to look for a lost dog reaches a surveillance centre. The aerial robot takes off and starts a mission to find the lost dog around the arena. During the flight, the robot must detect and avoid obstacles. The dog must be detected and its position estimated by the robotic system, and reported to the surveillance centre, along with live video feed. Once the dog is found, the robot can go back to the take-off location. Meanwhile looking for the dog, if any of the two other objects (i.e. plastic bag or cardboard box) is detected, its position has to be reported to the MK:DataHub.

DH interaction. Type: data generation. The aerial robot should update the MK:DataHub with the location of the dog continuously, as well as of the other two objects.

Main functionality(ies). The main functionality tested is Object Perception.

Auxiliary functionalities. Other required functionality is Indoor Navigation.


  • Detecting the dog and reporting its correct location.
  • Detecting the other objects and reporting their correct location.
  • Transmitting live images/video to the MK:DataHub.
  • Landing at the starting location.

Penalising behaviours.

  • Manual interventions to the aerial robot in case it needs to be recovered from a failure.
  • Hitting any of the structures in the flying arena (obstacles, safety net…).
  • Damaging the structures in the flying arena (obstacles, safety net…).
  • Hitting the dog.
  • Leaving the flying arena.
  • Failing in transmitting data to the MK:DataHub.