Boston Dynamics' impressive robots, capable of jumping, dancing, or opening doors, seem straight out of a science fiction movie.
Yet behind their physical prowess lies a more nuanced reality: although agile, these robots remain surprisingly limited in terms of intelligence. They perform preprogrammed tasks but still struggle to react autonomously to a changing environment.
A major current challenge in robotics is to integrate artificial intelligence systems into these machines so that they become capable of adapting and making decisions, transforming them into truly versatile assistants for humans.
The challenge is to design intelligent robots that can effectively interpret cues from their environment in order to adapt to real-world conditions, which are full of uncertainties (such as sidewalks and uneven ground) and variations (such as weather conditions or high or low passenger traffic at a train station…).
In industry, robots are particularly numerous compared to other sectors, but they are far from being well-suited to changing environments because they are limited by their perceptual capabilities and their ability to generate new behaviors in response to unforeseen situations.
We believe it would be interesting to design a human-robot collaboration based on an understanding of human capabilities—which the robot would complement.
By announcing the transition to Industry 5.0, the European Commission aims to improve industrial productivity and workers’ well-being on the job. That is why it is proposing the use of collaborative robots (also known as “cobots”) to assist humans in their work.
Indeed, there has been a growing number of robots being deployed in industry. In fact, the deployment of robots is increasing by 5% each year. However, these robots can have negative impacts on workers when technological development has driven innovation while ignoring the actual needs of the workplace. For example, a study has shown that experts in their field become frustrated and have difficulty accepting technology that is supposed to “help” them when it is not tailored to their needs.
Today, it seems essential to focus cobot design on the actual needs and constraints of the people who will be working alongside robots in their workplaces.
This involves understanding the work “without the cobot” and then working with employees to identify the tasks where the cobot could provide assistance. Adapting cobots to the needs of humans at work involves two approaches. The first requires adjusting the robot’s physical capabilities: its movements (trajectory, speed) and the tools attached to it (grippers, suction cups).
The second step should enable the cobot to enter a reasoning loop similar to that of a human: this loop begins with the perception of the environment, and then the information gathered from the real world is analyzed by the algorithm that controls the cobot’s behavior.
This is where artificial intelligence (AI) systems come into play.
As part of the Marvin team at the Grenoble Computer Science Laboratory, we are developing AI techniques specifically designed to take into account the physical environment of robots and enable them to reason and make decisions autonomously.
These artificial intelligence algorithms are based on a prior definition of the actions the robot can perform, the constraints that must be met for their execution, the state of the perceived world, and the target to be achieved.
The main challenge lies in the large number of possible actions, even for seemingly simple problems. Sensors are used to collect real-time data about the cobot’s environment. With constraints such as “preserving human initiative” or “preventing humans from being exposed to hazardous materials” guiding robotic decision-making, algorithms enable the robot to continuously recalculate the “actions to be performed” by taking into account changes in the real world and analyzing their consequences.
Our experiments aim to analyze the impact of cobotic collaboration (a cobot working alongside a human) during an industrial-style assembly task (participants are given a model and assembly parts that they must replicate with the cobot as quickly as possible).
The results of our experiments show that human-robot collaboration in the workplace can be beneficial overall for humans and their performance. In fact, we show that the human’s workload (both physical and mental) remains stable when performing the task with a collaborative robot, whereas this is not the case when the human performs the task alone or with another human. Interestingly, participants—regardless of whether they had previously worked with a cobot in another context—had a high level of trust in the cobot and expected the collaboration to be enjoyable.
This type of system would also benefit the company. In fact, human-cobot collaboration leads to better task performance than no collaboration or human-to-human collaboration; the reliability and quality of work are enhanced in the case of human-cobot collaboration. In our experiments, operators are also exposed to fewer risks (they come into contact with fewer elements classified as hazardous) when the cobot adapts, thanks to AI, to the safety constraints imposed by the task.
On the other hand, there is one consistent negative effect across all our experiments: an increase in the time it takes to complete the task. In fact, participants always take longer to complete the task when they collaborate with a cobot than when they collaborate with a human (or when they do not collaborate at all).
This finding is somewhat surprising and raises some fundamental questions: Could the development of robotic collaboration enhance workplace efficiency? Our study suggests that this may not be the case with current cobots, but that they do promote the quality of work and human well-being.
In conclusion, it is essential to design AI-powered cobots that take into account human needs and constraints and that reason based on data derived from actual observations of humans. Thus, the design of cobots requires, above all, close collaboration between researchers in computer science, robotics, and occupational psychology and ergonomics, with the goal of working together to improve working conditions.
The “Investissement d’avenir” program ( ANR-15-IDEX-02 CDP BOOT ) and the PACBOT project were supported by the National Research Agency (ANR), which funds project-based research in France. Its mission is to support and promote the development of basic and applied research across all disciplines, and to strengthen the dialogue between science and society. For more information, visit theANR website.
This article is republished from The Conversation under a Creative Commons license. Readthe original article.
