By Rafael Jiménez Sánchez, Senior EOD Advisor at aunav, by everis ADS
A revolution occurs when the entire society, or part of it, evolves, in a short period of time, into a new model. However, a true revolution occurs when the entire society is transformed for good, and right now, it is changing at an incredible pace. In this article I try to provide some clues to improve our response to hazards using previous experience.
Innovation and transformation are the trends shaping today’s world. We can put experience to work by exploring new ways of doing things and avoiding repeating the same mistakes time and again. We could use human behaviour and experience to find out where, when and how the problems occur and, consequently, design tools to overcome them. This approach is sound and simple in theory. The conflicts in Afghanistan and Iraq, the terrorist threat, the COVID-19 pandemic, etc., are clear examples of our limitations to successfully cope with VUCA environments (Volatility, Uncertainty, Complexity and Ambiguity).
Innovation and IT transformation are some of the new narratives that shape our society, but there are others. Adaptability and resilience also help us to overcome our shortcomings when facing VUCA environments.
The Revolution of Military Affairs (RMA) can be viewed as the military version of the IT concept. In addition to IT, RMA can be regarded from two different points of view: first, changes associated with technology having an impact on warfare; second, a revolution that results from the evolution of human society. War is no longer a fight for land and resources, and controlling hearts and minds can be more efficient in achieving the ultimate goals in any conflict. The “deep” RMA includes changes to the goals, organisational structure, strategy, ideology, and training of the army. Although RMA is beyond the scope of the article, we can use the concept as a starting point to tackle problems, in our case the design of Unmanned Ground Vehicles (UGV). Two possible guides, better combined, are suggested:
- Innovation and transformation
- Adaptation and resilience
FIRST STEP – UNDERSTANDING THE PROBLEM: COMPLEX THREATS
In the last century there has been an evolution of the hazards threatening us. These hazards, which should be controlled through human intervention, are expanding at the pace of technological evolution. Technology offers solutions that are right there, off the shelf, waiting for a practical application.
Is it possible to develop a common solution for the whole range of threats? Well, there is no “silver bullet”, that is what experts say. I can add that, regardless of the solution we choose, it has to face a dilemma: to cope with “business as usual” situations, or to tackle “black swans”. The A, B, C, D, E, F hazards appear in both situations: they may be expected or predicted, or just come out of the blue.
THE MORAL ISSUE
In 2017 a large group of leaders in the field of robotics signed an open letter calling upon the United Nations to prevent a global arms race in Lethal Autonomous Weapons (LAW). Artificial intelligence (AI) researchers are concerned about the collaboration between science and technology, and about those providing solutions to the requirements of the military. This cooperation could lead to the development of “killer robots”. It is odd that, for the first time, the International Humanitarian Law (IHL) is proposing a solution to this problem to come.
If the ultimate goal is to leave humans out of the equation, it would be the beginning of the end for us all. If a robot is designed to learn, it will eventually follow the criteria of the machine and disregard the human factor. Even if it follows a human model approach in its decisions, they (the robots) could fail, same as we (humans) do, precisely because it would be mirroring imperfect beings.
For the sake of our future, we should limit robotic work to dull, dirty and dangerous tasks, and never put a robot behind the trigger. The EOD task force provides a good example of why it is important to solve this dilemma, because the EOD UGV may carry out the functions of a second operator, but never that of the team leader. A similar example is the crew of an airplane. In the event of total human failure, a dead-man protocol has to be developed. This protocol is necessary before expanding the application of AI to warfare. Such a scenario brings to mind Kubrick’s film “Dr. Strangelove”.
THE CONCEPT ISSUE
Right now, we have three approaches to respond to hazardous situations or threats, and in all of them the human factor is fundamental. So, before starting the design of the UGV, human interaction with the unmanned system is something to be considered, together with a good understanding of the hazards. The second common factor of these approaches is the knowledge required to carry out the task of defeating a threat (preparation, Mine Risk Education (MRE), training).
In short, when adapting UGVs to tasks in hazardous environments, the starting point is to keep in mind the human factor and understand the impact of knowledge, in the sense of understanding the threat and putting into practice the methods to defeat it.
THE RESPONSE: KNOWLEDGE, TOOLS AND TRAINING
Simple things, such as learning, are not easy. We have to understand the fundamentals and disregard the fancy stuff or, in other words, seek the right approach and the right attitude. When I was responsible for the training of EOD-EOR specialists, I always insisted that making a mistake is not a bad thing, unless there is no room for a second try; in training, mistakes are a blessing in disguise because we can learn better for real-life situations.
Learning to cope with “uncommon” situations is a skill deeply rooted in our early experiences. To make this approach understandable, I use the “Ikea Furniture Model” when explaining the “training method”: there are two ways of assembling a chest of drawers, table, etc. The regular logic way: you open the box, look for the assembly instructions, read them, then look for the tools, and follow the step-by-step checklist. The second way is the typical Spanish model: you open the box, grab the instructions written in several languages, sometimes with an odd translation, then put it away as an unnecessary item, spread the content on the floor and begin the adventure of matching all parts, nuts, bolts and planks using your hands as the fool’s hammer and your intuition, normally under the scrutiny of your kin (who may contribute to add some stress to the task). The second method is apparently the worst, but the best if you want to develop a natural skill for improvisation and for finding solutions to unexpected problems.
The clever combination of both methods is the foundation of the Problem Based Learning Method (PBL). The key to PBL is to “find” the right problem and then set up a situation to work with it and evaluate the results. In the case of UGV this can be better explained with a list of ten problems. These ten items can be transformed into an obstacle course for the human-robot team. The closer they are to the real situation, the better. AI can be a good tool to improve the autonomous performance of unmanned systems.
- SELF-RECOVERY TASK. All UGVs are prone to fall in ditches or roll upside down in transversal slopes, or when negotiating obstacles. The quick recovery (hook and line) is an option, but it is not as good as the robot rescuing itself with a self-righting mechanism (more commonly referred to as Srimech, derived from the words ‘self-righting mechanism’), a device used by robots to flip themselves back onto their wheels if they get flipped over or stuck.
- MOBILITY. (Beyond Line Of Sight (BLOS) movement, corridors, confined spaces, slopes, bushes, etc.). If we try to achieve full mobility, then we have to design a tank. A tethered drone can be the substitute for mobility through “impossible areas”, but if the robot has to move in narrow spaces, variable geometry (a variable geometry robot is able to change its physical configuration during operation) to accommodate its width to the required situation is a feasible option.
- RUBBLE (Physical) AND EM (noise and jamming). Electromagnetic Compatibility (EMC) is desirable but, as mobility, it is difficult to achieve. Fiber optic guidance, or autonomous response can help in this case. A difficult situation is movement through rubble. Often, neither the operator nor the UGV have a full view of the obstacle. Again, a tethered drone can be a solution. Fiber optics, or an encoded laser beam to override the Remote Control (RC), could be desirable in environments saturated by EM radiation. There are several strategies to move through rubble. One is to keep the platform as horizontal as possible, which is called Horizontal Platform. Another is to try to maintain the centre of gravity within the area of the base of the platform, which is known as Stable CoG (e.g. the gripper lifts a heavy weight in a lateral position and the robot tilts the platform to balance this weight).
- ACCESS: STAIRS AND DOORS. This is a task where the standard robot has serious problems. The physical dimension is a constraint, and the need to use tools to negotiate obstacles may slow down the response. Again, a simple solution: a smaller tethered robot or drone could fill the gap to observe BLOS of the robot’s cameras. There are also remote-control strategies for autonomous stairclimbing.
5. DEEP SEARCH. One difficult job for the robot is to find a hidden threat, for example a buried mine, or to look inside a suspicious container using X-ray, or Non-Lineal Junction Detection (NLJD). Unlike humans, robots have no health constraints when using certain equipment. Robots with two arms allow the distance between the X-ray emitter and receiver to be dynamically varied, compared to static supports that require pre-configuration for operation.
6. SOFT MANIPULATION. Would the robot be able to open a matchbox and light a match? Swabbing and taking samples (gas, liquid, solid) require dexterity and repetitive routine protocols without the need for human control. Robots with two arms are paradigmatic to easily accomplish these tasks.
7. HARD MANIPULATION. Breaking a barrier, removing a dead body, or evacuating a casualty from a combat area to a first aid station are not routine tasks for an EOD robot, but these are critical scenarios. The UGV has to boost physical human strength. The concept of Search-Soft- Hard manipulation calls for adaptability (two arms/ one arm + tethering/master-slave/tools bay, etc.).
8. TOW AND TRANSPORT. Sometimes we forget that once the threat has been defeated, we have to continue to achieve the ultimate disposal. The robotic solution (with a towed container, for example) will maintain the operator out of harm’s way. Fire extinguishers, sprinklers, etc., may be necessary and, due to this, a towed container, or a cargo bay is necessary.
9. FAIL-SAFE CONTROL. Safe detonation is one of the tasks that will require a strict protocol and human-robot interaction. We take for granted that, with the standard EOD robot, this can be done. In reality, it is not that easy. The integration of a RC exploder is a solution. A coded laser LOS control could be desirable in hazardous situations.
10. DEAD-MAN CONTROL. Here is where artificial intelligence can play a significant role. Imagine that the robot has to extinguish a fire in a space with smoke, high temperatures and beyond RC reach. Or if the robot loses the communication link with the OCU, it has automatically to return to the point where it can be recovered.
Well, these ten situations or scenarios are helpful to measure the performance of our robot, but at the same time they offer a pattern for the human-robot team to interact and learn. No system will be practical without an attached simulation suite, and no real learning is possible without realistic PBL. We have to learn how to assemble “furniture” with and without instructions.
In summary, I can say that the design of the “perfect UGV” is an impossible task, but if we stick to the fundamentals and use the PBL approach, the solution is possible, without obsessing about innovation but focusing on the interaction with the user and the adaptation of what is already proven to work in order to do a better job. Remember: fail, try again, and fail better. ■
ABOUT THE AUTHOR
Rafael Jiménez Sánchez joined the Spanish Military in 1978. After getting his commission as an Engineer Officer he was posted to the Mountain Brigade (Engineer Battalion) in San Sebastian. He graduated from the University of Zaragoza and got a Master Degree in Security and Defense. His Military records include courses such as UN Military Observer, Signal Officer, Camp Construction and EOD Officer.
His assignments and tours of duty since 1982 include Engineer Battalion at the Mountain Brigade, Instructor (Military Academy and Engineer School), Bridge Regiment, CID and EOD School. Tours of duty in Bosnia (1994-1996-1998), Instructor at ENTEC (2000), Chief Ops-Int and Spanish Engineer Unit Kabul (ISAF- 2002/2003).
Promoted to Colonel in May 2011. He was appointed Director of the International Demining Centre (CID) and EOD School (Spanish Army). Joined the Army reserve in 2018, since then is working as associated professor and EOD advisor for aunav robots, by everis Aerospace, Defense and Security.
Download PDF version of this article: Rafael Jiménez Sánchez – Senior EOD Advisor at aunav, by everis ADS – C-IED REPORT, Spring-Summer 2020
Counter-IED Report, Spring/Summer 2020