Rafael has developed and tested a new algorithm for its Spice-250 precision-guided standoff weapon that employs artificial intelligence and deep-learning technology to provide automatic target recognition capability. This adds to Spice-250’s automatic target acquisition and moving target homing capabilities.

Using AI and deep-learning technologies, the Spice’s electro-optic scene-matching algorithms can “learn” target characteristics before the strike. As the launch aircraft approaches the weapon release point, the pilot assigns a target type to the Spice weapon, which is then released from standoff distances of up to 75 km (40 nm). During the fly-out the Spice guides by inertial navigation, but as it approaches the target area the electro-optic sensor scans the scene, the advanced ATR functions detecting and recognizing the allocated target before attacking it with great precision.

This process can be conducted autonomously or with a human in the loop and has no reliance on GPS. Individual Spice-250 weapons can be targeted and launched independently against an array of targets. Two warhead options are available—general-purpose and penetrating—weighing 75 kg (165 pounds). The Spice-250 is light enough to be carried singly by relatively small aircraft, while a Smart Quad Rack allows the carriage of four Spice-250s from a single weapons pylon.