Within the frameworks of a grant of the Russian Science Foundation (RSF), researchers from South Ural State University are engineering a smart system for controlling the liftoff, flight and landing of unmanned aerial vehicles (UAV) with high-lift devices, which could also be used for coordination of the operation of blades in wind-driven power-plants. This system is based on artificial intelligence with a self-training neural network.
Our university’s researchers aim at reducing the consumption of fuel or battery power by an unmanned aerial vehicle, as well as at making the automatic stabilization in the air possible. In the process of designing, the scientists from Chelyabinsk have also come up with an option of wing adjustment for fixed-wing drones. They included under wings and top wings (which earlier had only been used in big aircraft) in the wing structure. These elements regulate the liftoff, landing, rotation angle, as well as the steadiness and stability of the flight of an object.
Installing such parts in the UAV wings allows to control the flight from a ground base without the involvement of human beings. The operator only has to put in the flight coordinates and path, and the new system will help strictly follow those. This control system developed by our university’s scientists is based on self-learning artificial intelligence. Receiving the signal from the equipment, it can adjust the operation of the under wings and top wings in a remote mode; while at the same time choosing in the database the optimal angles of rotation and landing the apparatus. If gust of winds occur, the drone manoeuvrability will increase without the additional consumption of fuel or battery power since separate moving parts on the wings will make the process of the drone stabilization in the air easier. The new electronic control system will help to faster respond to the possible abnormal situations during the UAV flight and will not allow for its trajectory offset in space.
“There exist control systems for unmanned aerial vehicles, but those do not imply the use of such elements as under wings and top wings since no one has ever added such parts to drone wings before,” explains Konstantin Osintsev, Head of the SUSU Department of Industrial Thermal Power Engineering. “Our invention is unique not only because of the supplementary adjustment elements, but also thanks to the neural-network add-in for the electronic control system. We are creating a neural network, which will be accumulating a database resulting from the UAV’s errors during flight. That is, when a drone is moving without control on behalf of a human being, it repeats the same actions multiple times; and the neural network at this time collects all the errors in the movement of the apparatus (deviation from inertia, destabilization if facing an obstacle or due to weather conditions, etc.), which could occur during each previous tried-and-tested action. By collecting such a database, the neural network will be able to predict and immediately eliminate all possible deviations during the UAV’s movement.”
All the data collected by the neural network is stored in the electronic control base. The intelligent system, trained on hundreds of UAV flights beforehand, will “understand” how to act in case of an abnormal situation in the air by adjusting the direction angles of the rear flaps or under wings that stabilize the aircraft position in space.
In addition, the new neural-network-based control system will be able to independently select the optimal trajectories for the UAV liftoff or landing with consideration to the current meteorological conditions and the drone speed.
The main advantage of this unique control system is the fuel- or battery-power-saving consumption by UAV. This will allow to increase the claimed flying range at a certain speed by at least four percent.
Today, the SUSU scientists are engaged in producing the experimental elements of the high-lift devices (under wings and top wings) using a 3D printer. The researchers expose these to different effects from air flows in a wind tunnel at various speeds and elevation angles.
This multi-purpose invention by the scientists from Chelyabinsk can also be used for protection of blades in wind-driven power-plants, which are installed in regions where the wind is blowing all the year round. Wind blasts often break the blades in wind-powered generators, and because of that the efficiency of the whole wind farm drops (power generation has to be suspended).
The same elements of the high-lift devices produced for UAVs can adjust the movement of the blades in wind-driven power-plants: they can immediately and completely stop the wind-powered generator itself in case of hurricanes, or in squall wind redirect the turbulent air flows in order to reduce their impact on the blades. Such an approach will allow to operate wind-driven power-plants even at gusts of wind of 20 m/s to 30 m/s without losses in the efficiency, as well as eliminate the equipment downtime.
The moving elements in the high-lift devices installed in the blades will be controlled by the same neural-network-based system without the routine involvement of the operator.