Autonomous Weapon Systems are robotic weapons that can sense and act unilaterally depending upon how they are programmed. They select targets and deliver lethality without human supervision. But material developments are required in hardware and software if this self-learning weapon category can legally be deployed on the battlefield. This book unpicks the manifest complexity involved in fielding a weapon that can operate without oversight while still retaining value as a battlefield asset. The subject's importance is that several well-tried concepts that have long comprised battlecraft may no longer be fit for purpose. The book dissects the highly coupled nature of routines that are envisaged for AWS operation in order to demonstrate why it will be enduringly problematic - and unwelcome - to remove humans from tomorrow's battlefields.