The use of AI in UAVs can be divided into two key areas:

• Emulation of a "manned aircraft" in terms of interventions that a pilot would apply for a manned aircraft. For example sense and avoid technology that deconflicts the two platforms, some these would be regarded as an extension of flying aids already available to human pilots such as ADS-B (Out) and ADS-B (In). See the extensive work done to fly UAVs through civil controlled airspace.
• Image/data assessment; auto image recognition is already a proven technnology (for example your phone will now automatically examine an image and "tag" a name from your previous photos. The "person in the loop" to examine for example video footage is very expensive and time consuming, however the use of image recognition software (working to a ruleset) linked to an appropriate AI would allow for the AI to identify "items of interest" for the person in the loop to further assess and make decisions.

Dear Mr. Amine Mohammed Taberkit,

In attachment you will find the paper, devoted to the problem, you are interesting in. The list of references also will be useful for you.

Best regards,

A. Tunik.

If you break down a UAV in different components such as low level controller, state estimation, navigation and high level applications (inspection, tracking…) you can applied AI, actually Deep Neural Net (DNN) and Deep Reinforcement Learning (DRL) to almost all of these components, namely :

Regarding low level controllers, DRL is often applied with the objective to replace PID controllers by a neural net [1][2] ;

Regarding navigation, there are lot of interesting research going on. The one that come to my mind are Supervised Deep Learning for outdoor navigation [3][4] and DRL for indoor navigation [5] ;

Regarding sensing, researchers are designing DNN to extract 3D information from a pair of cameras [6] or even a monocular camera [7]. The depth information is then used to plan a path;

Regarding high level applications, researchers are focusing on objects tracking – CNN for object recognition and RNN (Recurrent Neural Net) in order to “model” the movement of the target [8]. Another interesting research consist in tracking other MAV for instance in order to avoid collision in a swarm [9]. Another controversial research aims at detecting violent behavior in a croud [10].

I provided the references that came in my mind when I wrote the answer, but clearly you have a lot more papers on the aforementioned topics.

Hope that help

[1] Zhang, T., Kahn, G., Levine, S., & Abbeel, P. (2016). Learning deep control policies for autonomous aerial vehicles with MPC-guided policy search. 2016 IEEE International Conference on Robotics and Automation (ICRA), 528-535.

[2] Koch, W., Mancuso, R., & Bestavros, A. (2019). Neuroflight: Next Generation Flight Control Firmware. CoRR, abs/1901.06553.

[3] Loquercio, A., Maqueda, A.I., del-Blanco, C.R., & Scaramuzza, D. (2018). DroNet: Learning to Fly by Driving. IEEE Robotics and Automation Letters, 3, 1088-1095.

[4] Smolyanskiy, N., Kamenev, A., Smith, J., & Birchfield, S.T. (2017). Toward low-flying autonomous MAV trail navigation using deep neural networks for environmental awareness. 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 4241-4247.

[5] Sadeghi, F., & Levine, S. (2017). CAD2RL: Real Single-Image Flight Without a Single Real Image. CoRR, abs/1611.04201.

[6] Smolyanskiy, N., Kamenev, A., & Birchfield, S.T. (2018). On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 1120-11208.

[7] Wang, K., & Shen, S. (2018). MVDepthNet: Real-Time Multiview Depth Estimation Neural Network. 2018 International Conference on 3D Vision (3DV), 248-257.

[8] Ning, G., Zhang, Z., Huang, C., He, Z., Ren, X., & Wang, H. (2017). Spatially supervised recurrent convolutional neural networks for visual object tracking. 2017 IEEE International Symposium on Circuits and Systems (ISCAS), 1-4.

[9] Schilling, F., Lecoeur, J., Schiano, F., & Floreano, D. (2018). Learning Vision-based Cohesive Flight in Drone Swarms. CoRR, abs/1809.00543.

[10] Singh, A., Patil, D., & Omkar, S.N. (2018). Eye in the Sky: Real-Time Drone Surveillance System (DSS) for Violent Individuals Identification Using ScatterNet Hybrid Deep Learning Network. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 1710-17108.

Article Bio-Inspired Approaches for Energy-Efficient Localization an...

Image processing applications based on UAV can be a prominent research area. Check the following paper.

Article A Blockchain-Based Artificial Intelligence-Empowered Contagi...

Also, distributed data training handled by UAV is also another hot topic.

Article A Digital Twin-Based Drone-Assisted Secure Data Aggregation ...

Article FBI: A Federated Learning-Based Blockchain-Embedded Data Acc...

I recommend this article

https://www.mdpi.com/2504-446X/7/2/89