Cadets designing, building unmanned aerial vehicles

  • Published
  • By Eddie Kovsky
  • U.S. Air Force Academy Public Affairs

At the Sensor-based Intelligent Robotics Laboratory in the department of electrical and computer engineering here, cadets and faculty are designing and building more sophisticated autonomous unmanned aerial vehicles.

These UAVs will able to communicate with each other and search, track and destroy targets as a team. Another type will be able to hover in place.

The department is working on two separate, but interdependent, projects to develop a distributed network for a set of UAVs. For the first project, six cadets majoring in electrical engineering, computer engineering and system engineering management are working with faculty to build an air-to-air UAV communication network as their capstone senior design project.

The ongoing project, an extension of work done last year, is sponsored by the Air Force Office of Scientific Research. The research team must design, build and test a system of network made of one ground tracking station and three UAVs. The system will send and receive telemetry and video data within the network, detect and capture targets using controllable sensor systems and display multiple video images captured by onboard cameras on a ground tracking station screen.

This semester, cadets are performing design tradeoff studies for all sub-component systems as they prepare to adapt a network system for flight.

“Decisions have to be made based on contradicting factors,” said Dr. Daniel Pack, engineering professor. “Right now we’re focusing our efforts on the issues related to the system design before we build and test a prototype this spring.”

The network must allow all three UAVs to communicate with the ground tracking station, while also allowing each UAV to communicate with each other -- independent of the ground control. Bandwidth usage is only one of several considerations cadets have to account for as they make engineering decisions.

“The challenge is to have three video feeds transmitted simultaneously within the network using a limited bandwidth,” Dr. Pack said.

The problem with displaying video data on a ground tracking station is similar to the surveillance system the department of computer science has been working on this semester. But in this scenario, the multiple video signals may need to pass through all three linked UAVs in the air before the data is collected on the ground, where it is synchronized with the telemetry data.

Though the chain link is a strain on the bandwidth constraints, it allows the UAVs to cover a larger operation area away from the ground tracking station. Cadets are currently working on the communication system analysis portion of the project to determine the best video signal they can expect to send in a realistic case without destabilizing the network. They are also working to improve the performance of an onboard camera sensor system.

Cadets have designed their own gimbal -- a bracket that holds the camera mounted on the UAV -- that can rotate the camera in three dimensions. Rotating the gimbal in flight to maintain a target in the camera view is easier than flying the UAV over the target again. Using the sensor system, multiple UAVs could view a single object from multiple angles without ever losing sight of the target.

There are many other factors to consider in the design phase. The number and weight of the subsystems onboard decrease flight time. For example, the sensor system offers target detection capability, but adds weight and volume on a UAV. A fast onboard computer provides autonomous control capability, but uses more power, space, weight and costs more money.

Once the air-to-air network is operational, researchers are developing a way to use the chain link to control cooperative UAVs in flight. Eight faculty members in the department, along with outside researchers, are working to develop cooperative algorithms for multiple UAVs to search, detect and locate ground targets.

In general, compared to a single sophisticated UAV, a set can complete a task faster, collect more relevant data, more accurately sense the environment, provide redundancy -- if one UAV fails, the group can compensate -- and offer alternatives.

In many military applications, multiple UAVs are necessary to perform the required tasks, such as providing the close coverage of a large area and the capability to engage in multiple mission-critical activities at the same time. Such tasks are not feasible for a single system, no matter how powerful the system may be, the professor said.

The relatively low sophistication involved in creating a set of UAVs, compared to one complex UAV, makes multiple cooperative UAVs more attractive to researchers and operators, he said.

In addition to controlling and coordinating multiple UAVs, researchers also want the cooperative UAVs to be able to autonomously detect and destroy a simulated target on the ground. The ground targets are mobile radio signal emitters that can turn on and off with unpredictable durations. Because the autopilots can make their own decisions, one UAV can search an area while another can deliver a weapon once it discovers a target.

Cadets working on the earlier stage of the project last year successively built three mobile robots that could cooperatively detect and destroy a simulated target. The next step is to successfully complete the same demonstration with UAVs in flight.

For the third project, four cadets with mentors from electrical and computer engineering and the aeronautics departments are continuing work to design a hovering UAV in the senior capstone design course.

The aeronautics department delivered the platform for the vehicle. Engineering cadets are working toward a final flying system with onboard control electronic components that can autonomously lift, land and control its flight while avoiding obstacles. Last year’s prototype was able to fly for several seconds before it spun out of control and crashed.

The military is interested in the possibility of replacing human intelligence with hovering UAVs. A hovering UAV can conduct close surveillance on a specific area for long periods of time.

Unlike a standard UAV, a hovering UAV does not have to keep moving to monitor a stationary target. It can perch on a tree or a building and continuously collect important data.

The success of these projects can directly impact future Air Force use of UAVs, officials said.