This is how a robot learns to improve its movements to adapt to any type of terrain



How can a robot adapt to the surface on which it circulates? Here are some of the keys to this new bill technology.

Robotics has incorporated multiple solutions over the last few years. The improvement of the software systems has served to add new functions and, of course, to implement solutions to improve those aspects in which relevant advances had already been achieved. Following this commitment, the introduction of novelties has become one of the hallmarks of those companies who have been working for years on their projects.

Naturally, MIT or Boston Dynamics are 2 of the most recognizable bodies in this regard. And, within their respective research centers, it is striking to see how Spot and a cheetah are the most advanced alternatives. Each one in its respective field is capable of developing actions with great potential for the next few years. Now, what happens if we focus on one of the 2 models in particular? The Cheetah has incorporated important updates.

One of the premises for having robotic units that can be autonomous is their ability to adapt to each specific terrain. Based on this, the arrangement of surfaces with different levels of adherence makes artificial intelligence tries to adapt each one of them. After all, you can’t send the same force to each leg on steep terrain as on completely flat terrain.

For this reason, a continuous learning system has been created that has had as its main reference adaptation of the switchboard to each of the surfaces. Based on this, it is capable of managing all the information to favor the movements of each leg in particular. As you can see, this requires great agility on the part of the software system.

Let us see, therefore, what this new approach proposed by the company consists of, why we are faced with a differential solution in terms of development and, of course, what implications the availability of a technology of these characteristics will have for the coming years.



The robotic mini cheetah capable of adapting to all types of asphalt

The first thing that draws attention is to see how this model has managed to increase its maximum speed compared to any previous record. On this occasion, a record of 14.04 km/h has been set, which demonstrates the improvement in ability to move from one place to another. However, the fundamental thing in terms of productivity is based on the development of a solution directly related to its adaptive capacity.

In the previous images it can be seen how throughout the learning process an improvement in the robot’s ability to circulate between the stones is achieved. It starts off stumbling, while finally getting send the necessary force to its robotic legs to be able to move. It is, as you can see, a process that may have important implications over the next few months in terms of its performance.

One of the most striking aspects of the progress made is directly related to the change in surface. Thanks to the incorporation of continuous tests, it has been possible to introduce parameters for make possible a change of surface without slipping. This is, it could be said, the main advantage offered by the solution proposed by this robotic unit. After all, the future humanoid should be able to adapt to the terrain it travels on, right?

Important advances to achieve practical applications for these robots

As has been observed over the last few years, important advances have been made in this area. In the past we have been able see how operators introduced traps to Spot. We have also been able to verify how the movements of the human being have been copied in a remarkable way until they become very natural. The passage of time and, above all, the largest investments in the field, are causing significant improvements in robotics to be observed.

The robotic cheetah is capable of improving its own movements over time. Europe Press

The prototype that has been seen in the previous images corresponds to a small-scale model. The original version is already capable of drive at a speed of 45.5 km/h, which shows to what extent we are facing a differential technology in terms of improvements in its performance. Taking into account all these advances carried out in recent years, what can we expect for the future?

Everything seems to indicate that, one day, the human being will live with humanoids created by the same. Even so, we will have to wait to see under what format this coexistence takes place, since artificial intelligence programs still have to introduce a whole series of variables to be exploited.

Related topics: Technology



Related News

Apple’s car is favored by consumers before it even exists – 26% would consider buying one from an iPhone maker – as firm hires former VP of Hyundai’s autonomous vehicle lab

Apple's non-existent self-driving car is already being adopted by consumers as a new survey shows that 26 percent of drivers would 'definitely consider'

The great paradox of the gas cap: why we have come out ahead despite having to pay more than ever

What we save on one side, we pay for on the other. The cap on gas has helped lower the price of a megawatt hour of energy, but the additional compensation to

AMD Ryzen 5 7600X Stock and 5.6 GHz OC CPU Benchmark Leaked Out, Equivalent to Core i9-12900K in Single-Threaded Test

More benchmarks of AMD's Ryzen 5 7600X CPUs in stock and overclocked configurations have leaked to Bilibili.

Google Drive: the steps to share a file as a link from the app

google drive It is one of the most feasible options to store and send any type of file, photos, videos, among others. This because the apps It features

More than 1,800 iOS and Android apps are leaking your data

The alarm is being raised by Symantec, a cyber security organization with which the private information of millions of people can be accessed through multiple

Physicists break down the speed of light with pulses inside hot plasma

Most of us are familiar with the prevailing law that limits how quickly information can travel through empty space: the speed of light, which tops out at