Abstract: Natural selection makes survival of the fittest. During billions of years of evolution, animals have formed the survival and reproduction abilities in the extreme environment in the presence of various challenges (e.g., virus invasion, environmental upheaval, natural enemy invasion, and population competition). Similar to animals, unmanned systems are affected by kinds of uncertain and abnormal factors including disturbance, attack, denies, damage, and fault. In the disturbance and confrontation environments, enabling unmanned systems to "self-learning" or "survival" has become one challenging problem. From the perspective of control, this paper proposes the concept of survival intelligence and three related key control factors of unmanned systems:safety control, immunity control, and green control. Safety control refers to the anti-disturbance and fault-tolerant control of unmanned systems subject to multiple disturbances and faults. Immunity control includes the autonomous sensing, self-regulation, and learning evolution of the system in face of confrontation and competition. Green control implies the control for "energy-saving""time-saving""strength-saving" and "burden-saving" of the unmanned systems under multiple constraints. Systems are like life, and devices are like organs. The purpose is to give life to unmanned systems in terms of intelligent behaviors and functions, and to achieve the leap from "method design""system design" to "behavior design", so as to improve the survival ability of unmanned systems in dangerous, extreme, special, and harsh environments.

Key words: unmanned system, survival intelligence, safety control, immunity control, green control, disturbance absorption