AI Algorithm for Predicting and Optimizing Trajectory of UAV Swarm

TL;DR

The paper tackles safe, scalable UAV swarm trajectory prediction and collision avoidance using AI. It systematically compares a range of activation functions on a FFNN with a single hidden layer, introduces the AdaptoSwelliGauss activation for enhanced accuracy, and develops an ICDAB framework that integrates collision detection, avoidance, and batching. Key findings show AdaptoSwelliGauss achieves ultra-low $MSE$ values (e.g., $3.059\times10^{-14}$ for X) and robust $SMAPE$ (≈$0.12$–$0.14\%$), while ICDAB dramatically reduces collisions (from 262 to 0) and improves batching efficiency (7→5 batches, 38→315 max per batch). The work uses a 500-UAV dataset with 201 waypoints, trained over 1000 epochs, illustrating significant gains in trajectory accuracy and safety for autonomous UAV swarms and suggesting avenues for future enhancements with CNNs, LiDAR, and multi-agent reinforcement learning.

Abstract

This paper explores the application of Artificial Intelligence (AI) techniques for generating the trajectories of fleets of Unmanned Aerial Vehicles (UAVs). The two main challenges addressed include accurately predicting the paths of UAVs and efficiently avoiding collisions between them. Firstly, the paper systematically applies a diverse set of activation functions to a Feedforward Neural Network (FFNN) with a single hidden layer, which enhances the accuracy of the predicted path compared to previous work. Secondly, we introduce a novel activation function, AdaptoSwelliGauss, which is a sophisticated fusion of Swish and Elliott activations, seamlessly integrated with a scaled and shifted Gaussian component. Swish facilitates smooth transitions, Elliott captures abrupt trajectory changes, and the scaled and shifted Gaussian enhances robustness against noise. This dynamic combination is specifically designed to excel in capturing the complexities of UAV trajectory prediction. This new activation function gives substantially better accuracy than all existing activation functions. Thirdly, we propose a novel Integrated Collision Detection, Avoidance, and Batching (ICDAB) strategy that merges two complementary UAV collision avoidance techniques: changing UAV trajectories and altering their starting times, also referred to as batching. This integration helps overcome the disadvantages of both - reduction in the number of trajectory manipulations, which avoids overly convoluted paths in the first technique, and smaller batch sizes, which reduce overall takeoff time in the second.

Figures (6)

• Figure 1: FFNN Architecture.
• Figure 2: AdaptoSwelliGauss activation function.
• Figure 3: Collision sphere around the UAV.
• Figure 4: Collision avoidance between two UAVs.
• Figure 5: Dataset of 500 UAVs.