He does not bite: how to make industrial robots safe for workers

We have already said that the meeting of people and machines in production mainly ends in peace. It is somewhere more efficient to automate production tasks, but in many areas it is still more useful to rely on the animated proletariat, but in general, robots and people are working more closely at factories, working literally side by side. The question arises: how to make sure that cars do not accidentally harm a person? We tell the story of the problem in which the first blood has already been shed, as well as about technologies that prevent these unpleasant collisions.

Even during the life of the science fiction and futurist Isaac Asimov, who formulated the three laws of the peaceful coexistence of a robot and a man, a machine was found that violated them.

By the way, if someone forgot, the laws are simple:

1. A robot cannot harm a person or, through inaction, allow a person to be harmed.
2. The robot must obey all orders given by a person, except in cases where these orders are contrary to the First Law.
3. The robot must take care of its safety to the extent that it does not contradict the First or Second Laws.


The shorter the distance between humans and robots, the greater the risk of accidents. To ensure safety, careful management is required, including the separation of the robot and the work area. Source: Toshiba

On January 25, 1979, in the city of Flat Rock (Michigan, USA), Robert Williams, a 25-year-old Ford warehouse employee, was instructed to retrieve parts stored in a huge rack, which was also serviced by a five-level truck. It was developed by Litton Industries to move blanks in stock. Part of the car consisted of plain-colored vehicles for transportation — carts with rubber wheels equipped with mechanical manipulators for moving blanks. Williams was sent to do a job that the robot could not handle – some details were left without the attention of the machine. The worker climbed to the third level of the rack and began to carry out the task. In the meantime, one of the carts appeared here, which immediately killed the worker, who did not notice her approach, with a blow from the manipulator. Williams’ body remained on the shelf for half an hour until the workers found him. The robot routinely continued to move the blanks. Relatives of the worker sued $ 10 million in damage, and Williams, not even learning about it, went down in history as the first victim of the robot.

It is symbolic that the death of Williams happened on the anniversary of the premiere of Karel Chapek’s play “Rossum Universal Robots”, which took place in 1921. It was this work that gave us the concept of “robot”, as well as the first description of the total destruction of humanity by rebel machines.

However, to call the death of Robert Williams a murder would mean a great exaggeration, because the factory machine did not have the main “qualifying symptom” of intentional murder – the motive. And even in July 2016, when a police robot with explosives was used to eliminate a criminal in Dallas (Texas, USA), the actions of the machine were still led by a man.

Do robots often harm humans?

Although all robots, with the exception of the military, are designed so that none of Azimov’s rules are violated, victims cannot be completely avoided. Are there many of them? The statistics of dramatic incidents with robots cannot be called extensive, although injuries and even murders involving machines occur regularly.

According to a study by the Occupational Safety and Health Administration (OSHA), industrial robots caused at least 33 deaths and injuries in the workplace in that country over 30 years. However, there are more gloomy data. In 2013, German insurance companies calculated that about 100 incidents involving industrial robots occur in Germany every year.

In an earlier study in 1987, which covered enterprises from the USA, Germany, Sweden and Japan, it was found out exactly how robots injured workers: in 56% of cases they inflicted penetrating wounds, in 44% – blows. Most accidents were caused by poor organization of the workplace (20 of the 32 incidents analyzed), and human errors provoked only 13 unpleasant situations.

There are statistics about the unsuccessful use of robots in individual areas of their application. For example, as doctors say, medical cobots “opened their cemetery”: in 2013, a team of scientists analyzed statistics from the US Food and Drug Administration (FDA) and found that since 2000 In 2013, during surgical operations assisted by robots, 144 deaths, 1,391 injuries and 8 thousand device malfunctions occurred. Among them, two deaths and 52 injuries were caused by spontaneous shutdown of the robot during an operation or in the wrong movement. One death and 119 injuries occurred due to the fact that the elements of the robot or the equipment held by it fell on the patient.

Robots in surgery are used where the accuracy of the hands and eyes of even the most experienced surgeon is not enough – from spinal surgery to hair transplantation. Source: IEEE Spectrum YouTube Channel

Why have robots become more dangerous?

The main and almost philosophical problem: robots do not recognize themselves as part of the world, and therefore, can be dangerous to others. However, if 20-40 years ago the robot was an electromechanical machine with a very limited range of repetitive tasks, and it could literally be “kept in a cage”, that is, limited access to it so that it would not harm anyone by accident, today the situation is changing. In order for robots to replace individual workers or supplement their capabilities with their own, machines have to be released from cages. For one conveyor, two sorters can work simultaneously – live and mechanical. Moreover, robots become multi-tasking, for which they need to move around the workshop. Accordingly, a single working space of a person and a machine is formed, in which anything can happen.

Industrial robots present several types of threats depending on their origin:

• mechanical hazards that arise as a result of unintentional and unexpected movements or the robot losing tools;
• danger of electric shock, for example, contacts with live parts or connections;
• thermal hazards, such as those associated with hot surfaces or exposure to extreme temperatures;
• noise that can harm your hearing.

What provokes these threats? According to the observations of OSHA experts, many incidents involving robots do not occur under normal operating conditions, but in emergency situations, for example, during reprogramming, maintenance, repair, testing, tuning or adjustment. The following are external factors that are not dependent on the person – natural or technical, say, a power failure.

In total, it is customary to single out seven main reasons for robots getting out of control:

1) control errors, that is, errors in the control system or software, which leads to unstable behavior or an increase in the dangerous energy potential of the machine;
2) unauthorized access – violation by an untrained specialist of a security zone near the machine;
3) mechanical malfunctions – the most unpredictable and dangerous failures that can lead to improper or unexpected operation of the robot;
4) natural factors – in this group of reasons, everything that can affect the behavior of the robot due to natural causes, in particular electromagnetic or radio frequency interference, as well as adverse weather conditions;
5) power system failure – for example, pneumatic, hydraulic or electric drives can interfere with electrical signals in control lines; the result is a surge of energy, electric shock, an increased risk of fire, especially when something sparks in robots using combustible hydraulic oil;
6) improper installation of the robot or its elements provokes a lot of emergency situations, including when trying to correct errors;
7) the human factor: programming, pairing, control errors, safety violations.

In January 2019, the promoter of one of the shopping centers in Las Vegas got lost in the parking area and was shot down by the Tesla Model S. Although the promoter was in the wrong place, the driver of the car who turned off the autopilot was to blame. Source: Promobot YouTube Channel

Most often, emergency situations arise either when the human robot intervenes in the work of the robot, or due to changes in the environment, to which the mechanical worker cannot react. What technologies can prevent this?

What technologies ensure the safety of robots?

Three key skills are made a safe employee of the robot: 1) controlled stop, 2) control of speed and separation of zones, 3) limitation of power and strength. Explain these concepts.

Controlled Stop (Safety-rated Monitored Stop, SMS)

A controlled stop of the robot occurs as soon as a person is in a certain area near the machine, for example, to replace equipment, adjustments, settings or other actions directly with the robot. In this case, the power supply continues, and the robot automatically goes into idle mode. As soon as the operator leaves the control zone, the robot resumes operations without additional commands.
To measure the distance from the robot to a person (or other interference), remoteness sensors of various types (optical, sound, etc.) are used, which determine the distance to the object, sending a signal and receiving a response. To ensure maximum safety, the sensors have two parallel systems for transmitting and processing signals about the appearance of interference in a given zone. The signals arrive in two different modules in the robot controller and are processed separately, with different algorithms. Then they go through cross-validation. Accordingly, if one of the channels fails for any reason, the robot will still stop.

This is how a controlled manipulator stop works. The Master of University of Antwerp used to create this system Kinect v2 from Microsoft. Source: Nick van Oosterwyck / YouTube

Speed ​​and Separation Monitoring (SSM)

This more sophisticated safety technology involves changing the behavior of the robot when a person is in a certain area near the machine – for example, slowing down the robot.

Technically, this happens like this: a machine continuously measures the position and speed of an object in the field of visibility. SSM can be implemented in both static and dynamic conditions. The key to this is the technology of recognition of surrounding objects. In particular, Toshiba object recognition technology is applicable for SSM. Initially, we developed it to solve problems with warehouse robots, which were supposed to learn how to handle parcels, including different sizes, colors, weights and shapes. If a robot needs to see what it captures, then why not learn how to recognize colleagues, including humans? Toshiba recognition technology is based on the use of two cameras, an RGB camera and a 3D camera. Combining the pictures they received, the robot can determine the position of the object in space and the direction of its movement.

Power and Force Limited (PFL)

This is one of the most commonly used technologies, which allows you to avoid harm to humans when a person and machine come into contact. It involves the use of various techniques, including force-sensing sensors. It converts the measured components of the force and moment vectors into signals suitable for processing by a robot that “touches” certain objects. For example, such sensors are mounted on the brackets of Toshiba robots, which sort out different types of packages, including fragile ones. Before taking the package, the robot “feels” it and the sensor determines which object the machine deals with in shape, elasticity, size. Based on the data obtained, the robot selects the clamping force and the speed of handling the object.

What technologies will make the robot safer in the future?

Future industrial robots, unlike their predecessors, which were only capable of rhythmic repetition of a limited range of tasks, will interact with a changing environment and humans. In order for the robot to be able to do this and perform atypical tasks, it must, to one degree or another, master the basic human skills, in particular, listening.

More recently, Toshiba introduced a new technology for speech recognition based on artificial intelligence. This is the first such development in the world: any robot or simple electronic device will be able to process voice commands without connecting to the Internet and cloud data. In other words, the processing device is integrated into the machine.

First, the neural network processes the sound, separating voice commands from extraneous noise. Next, a technique for expanding data in a neural network is used. Data extension is a method of studying small amounts of information, in this case, oral statements.


By randomly assigning zero weight to the connections between the nodes of the neural network, it is possible to generate simulated speech information, as if the speaker spoke differently. Successful identification of people occurs thanks to the training of AI based on samples of their speech, which allows you to recognize specific speakers, even when only a small number of statements are available. Source: Toshiba

Toshiba has reduced the number of speech patterns required to such an extent that the new technology can recognize the user in just three sentences. Such a technology is excellent. fit for robots and cobots with which you do not need to be verbose. Anyway, bye.

In the future, artificial intelligence will help robots not only hear better, but also see, think and move, which over time will completely eliminate the barriers between human workers and their artificial colleagues.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *