Shoal presents Machine Learning paper at MODSIM 2023

MODSIM 2023 - Machine learning–enabled missile system characterisation

At the 25th International Congress on Modelling and Simulation, held in Darwin in July 2023, a team from Shoal Group presented a paper on Machine Learning. The presentation was titled ‘Machine learning–enabled missile system characterisation’ and explored the feasibility of characterising missile systems using a multilayer perceptron (MLP) neural network trained on simulated data. A simulated dataset of 8000 ballistic missile trajectories was generated using a custom developed Python script. This dataset was split into a testing and training subset and each trajectory (samples of speed, latitude, longitude, and altitude) was processed into an n × 1 dimensional vector.

It was demonstrated that the trained neural network could infer the relationship between a ballistic trajectory input and weapon system parameters output.

The authors of the paper were Thomas Larkin, Sarah Whitehouse, Ben Mashford and Tim Yeung.

Read their paper: 2023 MODSIM – Abstract – Machine Learning Enabled Missile System Characterisation.

View their presentation: 2023 MODSIM – Presentation – Machine Learning Enabled Missile System Characterisation.

At the same conference, Shoal also presented a paper titled ‘Understanding the unpredictability of a ballistic missile’.

Shoal has been supporting clients in Defence for more than two decades. For further insight into how Shoal uses Machine Learning, please contact us for a confidential discussion.