AI-assisted Digital Terrain System for an Advanced Jet Trainer

This work focuses on the design of an artificial intelligence (AI)-assisted Digital Terrain System (DTS) for an advanced jet trainer. The DTS employs an advanced terrain server, a digital elevation map, and an efficient line-of-sight algorithm. The system utilizes a tuned search algorithm and three filter designs, including adaptive filters and a point mass filter, to adaptively trade-off between navigation accuracy and operating efficiency. Search algorithms such as TERCOM are known to be very compute intensive, running it continuously would require specialized or dedicated hardware on board. The AI model has the ability to independently switch between the search algorithm and bank of filters preserving navigation accuracy, while enabling real-time onboard implementation. Monte Carlo simulations shows that the proposed DTS can enable the low flying advanced jet trainer to perform terrain aided navigation without imposing new system requirements.