Military robots have existed in one way or another for well over a hundred years. The earliest examples include mines, torpedoes, and early-guided munitions. More ambitious systems have been developed in the last 80 years, and early attempts such as the Kettering Bug, Vergeltungswaffen-1(V-1 Buzzbomb), V-2 ballistic missile, Fritz X andHS-293 demonstrated the rapid evolution of inertial navigation systems (INS) that used pendulums, balances or gyroscopes for vehicle control. These vehicles followed an altitude-attitude-time program with altitude (or depth) pressure sensors and internal timers.
However, none of these systems were very precise. The situation has changed in these past two decades, as military robotics have become more powerful with unmanned air vehicles (UAVs), unmanned ground vehicles (UGVs), unmanned under-water vehicles (UUVs) and unmanned surface vehicles (USVs).
There are two main reasons behind the the adoption of robots in military, such as
- Standoff — we want to keep humans out of harm’s way.
- Precision — we want reliable and precise operation.
These are the key technologies, driving and supporting the current modern military robotics and robotics research.
- Electronic miniaturization
- Global Positioning
Electronic miniaturization has led to a revolution in robot sensing in the last decade as image sensors, lidar, radio detection and range (radar), and stereo imagers collapsed in size and cost. Miniaturization has also greatly changed telecommunications, providing a wide range of options from long-range (100 kilometers or less) and low bandwidth (100 kilobytes per second or less), short-range (1 kilometer or less) and high bandwidth Wi-Fi and cellular networks (50 megabytes per second or more) to iridium (2.4 kilobytes per second) beyond-line-of-sight. Virtually all robots require access to a Global Positioning System (GPS) that transmits accurate timing signals every second and provides a precise global coordinate and time synchronization system to build and execute models and plans. GPS has made the prospect of military robots at least conceivable for a wide range of low-complexity applications.
The following emerging technologies make these systems more reliable.
Emerging technologies in military robotics
- Probabilistic robotics
- Parallel processing
Over the past decade, probabilistic robotics has completely changed robotics’ face. The Sense-Model-Plan-Act (SMPA) cycle can be fragile, and one small error can be disastrous. Probabilistic robotics encompass techniques that can incorporate imperfect sensors into models, plans, or actions, and take uncertainty into account at every SMPA step, ensuring the plan’s outcome is best achieved. For example, by combining probabilistic tracking with new image processing, new vision algorithms can simultaneously locate and build models— a process known as simultaneous location and mapping (SLAM). This will lead to human-like sensing systems that can simultaneously locate and model.
For numerous reasons, the network, especially Internet, is an important robotics technology. Since the late 1970s, networking has rapidly become the Internet “cloud” in every facet of modern life. Network communication means robots can share sensing, modeling, and planning between robots. New standards promise interesting capabilities such as decentralized SMPA where sensing, modeling, planning and action could be distributed across multiple battlefield robots.
Networking has helped to some extent drive parallel processing. Parallel central processing unit architectures, ranging from multicore CPUs on the most desktops to graphics processors on gaming consoles, allow simultaneous processing to accelerate numerically intensive tasks, such as game rendering (modeling), Multi-Bot melee and search (planning), audio and gesture interfaces (sensing), all of which benefit future robotics.
The future of military robotics
Unmanned underwater, surface and air cars (UxV) will slowly enter complex environments in the foreseeable future as sensing, modeling, and planning improve.
UAVs go down to lower altitudes into more complex airspace as parallel processing and high-speed local networks deliver rapid sensor processing, shared models and faster structural flight planning. For example, shipboard operations, urban 3D mapping, and the clearance of organic convoys.
UUVs will penetrate closer to shore near ships and harbor facilities as parallel processing allow faster onboard sonar imaging for roles like harbor or hull inspection and waterway demining.
For complex operations, UGVs will need less handling, supported by powerful online parallel computing, long-term networking, and high mobility platforms.
Air-ground co-operation appears to be likely with overhead UAVs that provide top-down mapping and relay communication for ground vehicles. Examples are harder here, but probably include robots for squad support, indoor intelligence, ISR robots, and smart convoy vehicles.
Marsupial robots (robots carrying robots) are an important opportunity. This tactic enables a system to operate on several scales. For example, a high altitude, long-endurance, fixed-wing UAV could deliver smaller micro UAVs to provide both high altitude ISR and in-building mapping.