Copyright © 2025 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{179263,
        author = {Aditya Gupta and Rishav Goswami and Kapil and Aslam Kureshi},
        title = {LSTM-Based Neural Filter Framework for Oscillation Suppression in UAVs During Variable Payload Drop},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {6350-6355},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=179263},
        abstract = {Unmanned Aerial Vehicles (UAVs) experience severe vertical oscillations during payload drops due to sudden mass changes, leading to instability, excessive energy consumption, and safety risks in precision delivery tasks. This paper proposes a plug-in disturbance suppression framework that integrates a Long Short-Term Memory (LSTM) neural network with a Disturbance Observer (DOB) to mitigate oscillations without modifying the UAV’s proprietary baseline controller. The system inputs payload weight to predict disturbances via an LSTM, simulates the UAV’s response using a 1D linear time-invariant (LTI) model, and generates a corrective learning signal through an optimized filter, which is injected into the DOB before payload release. Stability is ensured through small gain theory and convex optimization. Experimental results on a SkyRiser 700 quadcopter demonstrate an 87% reduction in tracking error compared to no DOB, 47% versus standard DOB, and 35% versus adaptive PID, validating the framework’s efficacy and compatibility with commercial UAVs.},
        keywords = {Disturbance Observer, Long Short-Term Memory, Oscillation Suppression, Payload Drop, Unmanned Aerial Vehicles, Small Gain Theory, Adaptive PID},
        month = {May},
        }