Autonomous Robots in Tactical Role
Published on : Friday 07-08-2020
Autonomous robots meet the military and join the battlefield – Ravi Saini on integrating war strategy, data and machine learning.
Autonomous Robots are gaining momentum in the military and defence spheres. These include watercraft boats, Drones, UGV Tanks, Bipedal UGVs, Sentry UGVs and Autonomous Tanks, etc. These devices are utilised by militaries around the world for several decades. But with the developing/developed AI and Deep Learning technologies/algorithms, they have become more powerful for varieties of military functions and their importance is likely to grow exponentially in the very near future. According to a document prepared by the US Congressional Research Service (CRS), already AI has been incorporated into military operations in Iraq and Syria. China is actually leading the field of AI and its integration with robots – according a plan it released in 2017 – to capture the worldwide lead in AI development by 2030. Currently, China is primarily focused on using AI to make faster and more well-informed decisions, as well as on developing a variety of autonomous military vehicles. Russia is also active in military AI development, with a primary target on robotics 1
Today autonomous robots can be found performing various roles ranging from target tracking, target identification, search and rescue, explosive demilitarisation, reconnaissance, logistics support and, of course, lethal combat duties.
It is very critical getting information to soldiers during wars in real-time and along with saving lives, it helps deciding victory or defeat. What if AI-enabled robots were able to instantly provide information regarding the threat like their location, weapons and instantly sending crucial data to soldiers in combat?
AI based autonomous robots conduct long-duration data collection and data analysis, clean up environment contaminated by chemical weapons, or sweep routes for improvised explosive devices. In these roles, autonomous systems may reduce risk to soldiers engaged in combat and cut costs.
Robots powered with artificial intelligence can perform the following tasks:
Target recognition – AI in combination with other sensors can improve the ability of the system to locate the position of the target, automatically identifying the target and tracking them. Let's us say we deploy an autonomous robot (might be flying drone, or a moving robot, etc.), to the field. It can do the following tasks with great efficiency and accuracy:
a) Recognition of armed forces – It will identify if the forces/armed people are from enemy country or they are our own soldiers.
b) Weapon detection – It will recognising tanks and different weapons likes guns, bayonets, knives, bayonet-knife, grenades, handguns, less, lethal, rifles, carbines, shotguns, submachine guns, anti-tank/assault bombs, etc. The robot will inform about what weapons are being carried by the enemy.
c) Target tracking – It will track an object of interest which we want to target, e.g., it can track a specific tank.
Smart robots can be used for surveillance of border areas, identification and locating the potential threats, and transmitting these information to response teams and respective departments. This will help strengthen the security of military base camps as well as ensure the safety and efficiency of military personal at remote locations.
2. IED detection – It can be very difficult to detect IEDs (Improvised Explosive Devices) for humans, as they have wide variety of shapes and sizes, texture and orientation, and are typically camouflaged on the road and underneath the surface of the land. It is very easy to manufacture and hide them, which makes them a weapon of choice for terrorists across the globe.
Frequent surveillance of the high risk routes is critical to reduce casualties during military transports. Ground based patrolling is one of the most effective methods of surveillance. But at the same time it is not very feasible and effective method, because humans often find it difficult to concentrate on a task for a longer duration in an unknown constrained or unconstrained environment. Therefore, there is urgent need for robots that can assist military personnel in finding potential threats during driving.
We can develop a real-time early warning system/robot with decision-making capabilities. When mounted on a military vehicle, it can automatically detect the presence of those bombs by understanding suspicious changes in the environment with the help of video camera sensors, a GPS positioning system, and video/image-analytics.
3. Data information processing – Robots can use data skills to process large volumes of data from different sources to obtain valuable information.
We are living in the information age. Data is being generated and captured digitally by different digital means at tremendous volumes, at extremely fast pace, in ever-increasing velocities and in great varieties like text, images and videos, etc. Various devices on the robots such as mobile telephones, cameras, LEDs, and lot of other sensors all contribute to the exploding data volumes. This data can be browsed, stored, and shared, but its greatest value is when we can do efficient analysis of it. However, we might not have sufficient people to perform data analytics in order to derive actionable intelligence analysis.
4. AI conversational capabilities – Robots can communicate – are able to be in conversation with humans. Refer the following use case:
-Like we can ask about the current climate/weather condition where the device is located. Devices can use either external sensors or can be equipped with computer programs to get environmental data from the satellite.
-We can communicate with robot about our health condition and the device will assist us with the first aid.
In short there are enormous ways AI based robots can help while in conversation with humans. All these conversation capabilities can be done voice to voice using speech to text and text to speech technologies.
5. Location and mapping – The moving robots can be equipped with SLAM (simultaneous localisation and mapping) to be fully autonomous.
6. Person following capability – A robot can follow any specific person even if that person is standing in the crowd of, let's say, a hundred people.
7. Automatic graphical analysis – Data obtained from devices will be processed to do graphical analysis about the detected objects let's say percentage of weapons, percentage of persons, percentage of tanks detected by the camera. And send the report for human analyst attention.
8. Landmine detection and removal – The system will detect where are the landmines present using AI algorithms. The robot can be equipped with actuators to remove the land-mines or it will alert a human to remove it.
9. Monitoring soldiers’ physical and mental state – Biometrics aren’t just about identification of combatants. Sensors embedded in military uniforms and helmets can be used to send information to a command centre stating a soldier’s physical and mental condition, helping him or her survive otherwise dangerous enemy attacks. The physical health can be monitored with computer vision capabilities
10. Scene detection based shape-shifting capability – The recent natural and manmade disasters have advocated the study and research on the search and rescue autonomous robots. The shape shifting robots are modular in nature with the ability to change their shape and size, which gets changed depending on the nature of the environment analysed by the computer vision analytics.
11. Autonomy – Modern autonomous systems incorporate AI in some form or other, which makes them smarter and more intelligent. Autonomous systems are capable of supplementing and replacing humans, so they will be free for more complex and reasoning demanding work. In general, experts predict that the military is likely to gain significant benefits from autonomous systems by replacing humans in tasks that are ‘highly repetitive, dull, dangerous, or dirty’.
In conclusion, it may be noted that integration of various applications of AI and autonomy to create an eco-system for instant information sharing in war can change the diplomatic and strategic circumstances confronted by individual soldiers, massively improving prospects for survival.
Reference
1. https://fas.org/sgp/crs/natsec/R45178.pdf
Ravi Saini is working as a senior AI Engineer – Machine Learning, in China with total 5 years of experience, but has returned to India after the outbreak of Covid-19 pandemic. Holding a Masters degree in Robotics and Artificial Intelligence, Ravi is passionate about cutting-edge technologies and solving real world problems, he is currently building Applied Mathematics skills by taking several online courses such as Linear Algebra, Calculus, Complex Analysis, Statistics and Probabilities and use them on machine Learning, Data Science. Ravi is a practitioner in ML techniques, Deep Learning modules, Neural Networks, MARS and Big Data technologies such as Hadoop, Spark, HDFS and Kafka.
Ravi’s professional experience includes Proficiency in machine learning and deep learning skills for multiple applications including Computer Vision, Natural Language Processing. In his career Ravi has devolved various machine learning based solutions/prototypes for various industries including, medical, military, retail, sports, e-commerce, etc. One of his research papers has recently been selected by RESEARCH. The title of the paper is Design of Real Time Automated Deep Learning Algorithm based vector present classifier system and its control using bio-larvicide spray in urban/indoor environment. Ravi enjoys traveling, reading, swimming, sports and gym training.
