Mykhaylo Kyslynsky, CTO, LeoTronics
The latest popular and promising robotics trend is miniaturisation. The latest advances in bots and drones are based not on digital technology but on weight: the smaller the size of sensors, cameras and other intelligent equipment, the better.
Microrobotics is the field of miniature robotics, particularly mobile robots with a characteristic size of less than 1 mm. The terminology can also be used for robots capable of machining micrometre-sized components.
Microrobots were born from the advent of the microcontroller in the last decade of the 20th century and the beginning of miniature mechanical systems. The earliest research and conceptual development of such small robots occurred in the early 1970s in the (then) private study for the US intelligence agencies. The main goals of the research were to realise the need for organising prisoner rescue and electronic interception missions. At the time, essential miniaturisation support technologies were not fully developed, so no significant progress was made in prototype development.
Advances in wireless connectivity, especially Wi-Fi, have significantly increased the communication capability of microrobots and thus their capacity to coordinate with other microrobots to perform more complex tasks. Indeed, many recent studies have focused on microrobot communications.
Microrobot design features
Microrobots are less than 1 mm in size. They differ in this respect from millirobots, smaller than 10 cm (4 inches), small robots, smaller than 100 cm (39 inches), and nanorobots, at or below 1 μm or in a size range of 1 to 1,000 nm. Because of their small size, microrobots are potentially very cheap and can be used in large quantities (swarm robots) to explore environments that are too small for bigger robots or too dangerous for humans. Microrobots are especially useful when searching for survivors in collapsed buildings after an earthquake or scanning through the digestive tract. They lack physical and/or computational power, but this can be compensated for by using groups or swarms of them.
The way microrobots move depends on factors such as their purpose, makeup and size. One of the main hurdles is being able to move a microrobot using a minimal power source. Microrobots can use a small, lightweight battery source or can be powered by an external energy source, such as electromagnetism, ultrasound or light. Some microrobots in the medical field use biological motors, such as bacterial flagella, to obtain chemical energy from the surrounding fluid environment and propel themselves to their required destination.
To be autonomous, a microrobot must have the following four features. First, they should have sufficiently efficient micro- or nano-sensors.
Second, they should also have energy autonomy by means of efficient, low-power microbatteries or the ability to find and use an external energy source (solar, microwave, hydrogen, biomimetic ability to extract energy from organic matter, etc.). One way to save energy is to ensure that the various microrobot functions are activated only when needed and optimally. The rest of the time, they should be put on standby, which does not prevent them from moving passively (e.g., being carried by the wind, the current, a vehicle, etc.).
Third, they should have a built-in intelligent system (individual or collective in the case of robots with additional functions working together) and/or communication that allows interaction or remote control.
Fourth, they should have a curriculum complex enough to respond to the appearance of simple events and changes in the environment (stimuli) individually or collectively, with appropriate reactions.
Microrobot applications
Excellent prospects exist for microrobots in terms of tasks that for humans are, for example, hazardous, demand too high levels of accuracy and repeatability or impossible (e.g., needing to be undertaken in a too small space, a vacuum or airless environment). Industrial and technical robots are capable, for example, of making tiny parts or mechanisms, diagnosing or repairing the inside of a machine without it being disassembled and inspecting piping from the inside. Moreover, they can work in hazardous, vacuum or airless environments.
Microrobot applications in healthcare
For years, scientists have been developing ever smaller microrobots in attempts to deliver non-invasive, personalised medicine. The unit of measure for these microrobots is no longer a millimetre but a micron. They can diagnose and monitor diseases in real-time, measure blood sugar levels in diabetics and deliver a drug to a specific target in the body such as a tumour.
As mentioned previously, scientists have created microrobots that are capable of navigating through body fluids. They mimic the body's natural processes; they move forward using metachronal waves, similar to infusoria. Scientists are trying to find applications for these technologies, from treating cardiovascular disease to eye surgery. In the future, microrobots could remove plaque from arteries and break up kidney stones. Medical microrobots going about their jobs in bodies might one day be commonplace.
Microrobot applications in extreme conditions
Extreme conditions or environments are those classed as dangerous for people or technical devices, e.g., high radiation, aggressive chemicals, strong electromagnetic fields, high or low pressure and high or low temperature.
When robots perform various tasks at the scenes of natural or man-made disasters or military conflicts, the non-deterministic environment takes on special significance, where along with extreme external conditions, there is a high degree of uncertainty not only of the parameters of the environment itself but also of the operations to be performed. The resistance of robots to aggressive environmental conditions is ensured mainly by engineering and technical solutions. At the same time, the ability of robots to adequately and timely respond to unforeseen changes in ecological parameters depends primarily on the chosen control method, as well as the level of intelligence of the robot. Such robots must be equipped with a powerful onboard computing machine, an ample supply of onboard power resources and, often, a significant set of working organs. Such hoarding leads to increased dimensions and weight of a robot, which significantly limits the possible area of application.
Among the various tasks solved by extreme robotics, there are several for which a small-sized robot is desirable and sometimes necessary. These include surveillance of territories and water areas under conditions of organised enemy resistance, searching for victims in rubble after natural or man-made disasters, searching for and neutralising of explosive devices in anti-terrorist operations in dense urban settlements, and surface exploration of other planets.
Microrobots are required to undertake such tasks, but their small dimensions impose limitations such as:
- difficulty moving in an unprepared space because of relatively small protrusions and depressions;
- inability to move items, e.g., rock samples;
- small onboard energy reserve;
- small size and power consumption of communication means, leading to a restricted radio communication radius; and
- significantly limited number of working tools.
All the aforementioned limitations apply to a single microrobot. Therefore, an obvious solution might be to use a group of microrobots capable of combining their efforts. Microrobots can help each other to overcome obstacles and jointly carry out the transportation of a large item. Information exchange in a group of robots makes it possible to expand information about the environment available to each robot. At the same time, some tasks can be distributed between microrobots and performed in parallel. For example, while some robots in a group are collecting environmental data, others are collecting soil samples.
Moreover, group application of microrobots helps reduce the risk of task failure because the damage of one or several group microrobots in the general case (primarily in swarming and collective control methods) does not result in task failure, rather it reduces the efficiency of the group. On the other hand, damage to individual units of a single microrobot can disrupt the work it performs; and attempts to duplicate the most basic functional units of the robot lead to an increase in its mass, dimensions and cost, but do not increase its efficiency.
Methods for managing a microrobot group
The efficiency of microrobot groups depends mainly on the chosen control method, namely whether it is centralised or decentralised. In centralised control methods, a central control device has access to information about the state of all robots in the group and the environment. The control device evaluates the current situation and makes decisions about the actions of the group’s robots. The central control device may be located outside the group (e.g., on the operator's control panel) or onboard one of the robots in the group. In the latter case, one speaks of centralised control with a master.
Centralised control methods
Centralised control methods give good results when the number of robots in the group is small. As the group size increases, the load on the communication channel and the computational means of the control device increase. One solution is to use hierarchical control methods in which a group of robots is divided into subgroups, each with its leader (usually from among the group robots), and the subgroup leaders are controlled by a central control device located onboard one of the robots, or, more often, outside the group. However, hierarchical control methods complicate the nature of communications between robots in the group, making severe demands on onboard communication equipment. Interference in the communication channel is highly detrimental to the group operating under centralised control methods. In addition, the failure of a robot controlling robots in a group or subgroup means communication is lost and leads to severe problems.
Decentralised control methods
Decentralised control methods include collective, swarm control and swarm intelligence. In a collective control method, each robot in a group receives information from all other robots in the group and transmits the information it has collected about the environment and its state to the communication channel so that this information is available to all other robots in the group. Thus, information exchange in a group of robots under collective control is carried out according to the ‘each to all’ principle. This means that each robot can independently evaluate the current situation and decide on the necessary further actions.
Collective control methods allow the group to maintain operability in case of failure of one or several robots in that group. The load on the communication channel increases in direct proportion to the number of robots in the group. The limitation on onboard computing devices of robots also increases since they need to process the received information. Although the upper limit of permissible group size in collective control methods is, on average, higher than in centralised methods, the scalability of these methods leaves much to be desired.
In swarm control methods, there is no dedicated communication channel for information exchange between robots. Each robot collects information about the environment independently and independently decides on its subsequent actions to contribute to the group task. The absence of communication between robots in a group under swarm control methods allows for successfully solving of only those tasks that can be easily paralleled into independent unrelated subtasks. The main advantage of swarm control methods is scalability: the computational complexity of control tasks does not increase with increasing robot group size, allowing swarm methods to control huge groups of microrobots.
Swarm intelligence methods, already used for solving many practical problems, can be applied to control large groups of robots, which leads to the appearance of a separate direction, the so-called swarm robotics. Each group robot communicates only with some neighbouring robots falling within the visibility range limited by the range of its telecommunication devices (or limited artificially). Each robot makes its own decision about further actions based on simple local rules. The robot has access to the information about the environment that is collected on its own, as well as the information about the environment and the state of some robots in the group that is passed to it by neighbouring robots. The robot transmits the data collected about the environment and its form to the communication channel. This information becomes available to those robots whose line of sight this robot enters (in the case of the same radius of view, they are the same neighbouring robots).
This approach gives the robots more information about their environment than in swarm control methods. The information available concerns the area around them, i.e., is the most relevant. At the same time, scalability is preserved, since increasing the group size does not increase the load on the onboard computing devices. Thus, swarm intelligence methods offer excellent opportunities to develop mass-applicable microrobots, allowing the successful use of large groups of microrobots. However, the achievements of swarm robotics are limited to a few experimental projects; they are still almost uncommon in practice.
Obstacles in the development of microrobot group applications
One of the apparent obstacles to the development of swarm robotics is that multiple groups of microrobots must serve as control objects, which in turn implies their relatively inexpensive mass production. The progress in microelectronics, mechatronics and nanotechnology suggests that the mass production of microrobots will be possible and economically feasible shortly.
The second obstacle is the lack of general theory and approaches in creating and developing methods of swarm control in groups of robots. To date, much of the research has focused on using natural analogues of swarm intelligence methods to solve technical problems. Colonies of ants, swarms of bees, flocks of birds and schools of fish have served as models for various swarm intelligence methods.
Differences in the tasks and capabilities of natural and technical systems make it challenging to find and adapt natural algorithms to solve technical problems. Several types of research are devoted to creating artificial methods of swarm intelligence, initially intended to solve practical problems. Unfortunately, the lack of a unified approach complicates these studies. In fact, each new issue is solved almost from scratch each time. It is expected that using a suitable system to solve several tasks will significantly simplify the task of organising swarm interaction in groups of mass-applicable microrobots.
Small robots, significant prospects
Micro-level objects are governed by forces and laws that are very different from those at the macro level, so basic research into these phenomena seems inevitable. A significant challenge in using mobile microrobots is achieving predictable robot behaviour and controllability, including swarming behaviour. In addition, microrobots used for both military and civilian observation applications need reliable and long-term power sources. Such applications also imply significant challenges for microrobot design, depending on the application environment of flying, crawling or floating microrobots. Communication in microrobot swarms is also a big challenge in these applications. However, one thing is clear, and this is that while many microrobot projects and ideas are still in their infancy, this industry is poised for enormous growth because the opportunities are plentiful.
LeoTronics