New AI-Powered Missiles Could Be Harder to Dodge
Fighter jets and missiles just got smarter — and harder to escape
A new paper presents a look-angle-based model predictive control guidance (MPCG) method for missiles with strapdown seekers, which cannot directly measure line-of-sight rates. By using look angles and their derivatives as state variables, the method eliminates body-rate coupling and parasitic feedback. The guidance problem is solved as a nonlinear program with explicit field-of-view and acceleration constraints, while target acceleration is estimated using an adaptive extended Kalman filter with an interacting multiple model framework. Simulations against pitch/yaw weaving and barrel-roll maneuvers show MPCG delivers reliable interception and outperforms pure proportional navigation in stability and resilience, offering a practical solution for modern missile seekers with measurement limitations.
- New AI guidance system makes missiles better at tracking and hitting fast-moving targets like jets or drones
- Uses real-time prediction and adaptive filtering — more like a smart chess move than old-school straight-line tracking
- Raises concerns about arms races and the role of AI in military decisions
Why It Matters
AI may soon make missiles harder to dodge — changing who can win wars and how conflicts are fought.