Tag Archives: really
#439527 It’s (Still) Really Hard for Robots to ...
Every time we think that we’re getting a little bit closer to a household robot, new research comes out showing just how far we have to go. Certainly, we’ve seen lots of progress in specific areas like grasping and semantic understanding and whatnot, but putting it all together into a hardware platform that can actually get stuff done autonomously still seems quite a way off.
In a paper presented at ICRA 2021 this month, researchers from the University of Bremen conducted a “Robot Household Marathon Experiment,” where a PR2 robot was tasked with first setting a table for a simple breakfast and then cleaning up afterwards in order to “investigate and evaluate the scalability and the robustness aspects of mobile manipulation.” While this sort of thing kinda seems like something robots should have figured out, it may not surprise you to learn that it’s actually still a significant challenge.
PR2’s job here is to prepare breakfast by bringing a bowl, a spoon, a cup, a milk box, and a box of cereal to a dining table. After breakfast, the PR2 then has to place washable objects into the dishwasher, put the cereal box back into its storage location, toss the milk box into the trash. The objects vary in shape and appearance, and the robot is only given symbolic descriptions of object locations (in the fridge, on the counter). It’s a very realistic but also very challenging scenario, which probably explains why it takes the poor PR2 90 minutes to complete it.
First off, kudos to that PR2 for still doing solid robotics research, right? And this research is definitely solid—the fact that all of this stuff works as well as it does, perception, motion planning, grasping, high level strategizing, is incredibly impressive. Remember, this is 90 minutes of full autonomy doing tasks that are relatively complex in an environment that’s only semi-structured and somewhat, but not overly, robot-optimized. In fact, over five trials, the robot succeeded in the table setting task five times. It wasn’t flawless, and the PR2 did have particular trouble with grasping tricky objects like the spoon, but the framework that the researchers developed was able to successfully recover from every single failure by tweaking parameters and retrying the failed action. Arguably, failing a lot but also being able to recover a lot is even more useful than not failing at all, if you think long term.
The clean up task was more difficult for the PR2, and it suffered unrecoverable failures during two of the five trials. The paper describes what happened:
Cleaning the table was more challenging than table setting, due to the use of the dishwasher and the difficulty of sideways grasping objects located far away from the edge of the table. In two out of the five runs we encountered an unrecoverable failure. In one of the runs, due to the instability of the grasping trajectory and the robot not tracking it perfectly, the fingers of the robot ended up pushing the milk away during grasping, which resulted in a very unstable grasp. As a result, the box fell to the ground in the carrying phase. Although during the table setting the robot was able to pick up a toppled over cup and successfully bring it to the table, picking up the milk box from the ground was impossible for the PR2. The other unrecoverable failure was the dishwasher grid getting stuck in PR2’s finger. Another major failure happened when placing the cereal box into its vertical drawer, which was difficult because the robot had to reach very high and approach its joint limits. When the gripper opened, the box fell on a side in the shelf, which resulted in it being crushed when the drawer was closed.
Failure cases including unstably grasping the milk, getting stuck in the dishwasher, and crushing the cereal.
Photos: EASE
While we’re focusing a little bit on the failures here, that’s really just to illustrate the exceptionally challenging edge cases that the robot encountered. Again, I want to emphasize that while the PR2 was not successful all the time, its performance over 90 minutes of fully autonomous operation is still very impressive. And I really appreciate that the researchers committed to an experiment like this, putting their robot into a practical(ish) environment doing practical(ish) tasks under full autonomy over a long(ish) period of time. We often see lots of incremental research headed in this general direction, but it’ll take a lot more work like we’re seeing here for robots to get real-world useful enough to reliably handle those critical breakfast tasks.
The Robot Household Marathon Experiment, by Gayane Kazhoyan, Simon Stelter, Franklin Kenghagho Kenfack, Sebastian Koralewski and Michael Beetz from the CRC EASE at the Institute for Artificial Intelligence in Germany, was presented at ICRA 2021. Continue reading
#439327 It’s (Still) Really Hard for Robots to ...
Every time we think that we’re getting a little bit closer to a household robot, new research comes out showing just how far we have to go. Certainly, we’ve seen lots of progress in specific areas like grasping and semantic understanding and whatnot, but putting it all together into a hardware platform that can actually get stuff done autonomously still seems quite a way off.
In a paper presented at ICRA 2021 this month, researchers from the University of Bremen conducted a “Robot Household Marathon Experiment,” where a PR2 robot was tasked with first setting a table for a simple breakfast and then cleaning up afterwards in order to “investigate and evaluate the scalability and the robustness aspects of mobile manipulation.” While this sort of thing kinda seems like something robots should have figured out, it may not surprise you to learn that it’s actually still a significant challenge.
PR2’s job here is to prepare breakfast by bringing a bowl, a spoon, a cup, a milk box, and a box of cereal to a dining table. After breakfast, the PR2 then has to place washable objects into the dishwasher, put the cereal box back into its storage location, toss the milk box into the trash. The objects vary in shape and appearance, and the robot is only given symbolic descriptions of object locations (in the fridge, on the counter). It’s a very realistic but also very challenging scenario, which probably explains why it takes the poor PR2 90 minutes to complete it.
First off, kudos to that PR2 for still doing solid robotics research, right? And this research is definitely solid—the fact that all of this stuff works as well as it does, perception, motion planning, grasping, high level strategizing, is incredibly impressive. Remember, this is 90 minutes of full autonomy doing tasks that are relatively complex in an environment that’s only semi-structured and somewhat, but not overly, robot-optimized. In fact, over five trials, the robot succeeded in the table setting task five times. It wasn’t flawless, and the PR2 did have particular trouble with grasping tricky objects like the spoon, but the framework that the researchers developed was able to successfully recover from every single failure by tweaking parameters and retrying the failed action. Arguably, failing a lot but also being able to recover a lot is even more useful than not failing at all, if you think long term.
The clean up task was more difficult for the PR2, and it suffered unrecoverable failures during two of the five trials. The paper describes what happened:
Cleaning the table was more challenging than table setting, due to the use of the dishwasher and the difficulty of sideways grasping objects located far away from the edge of the table. In two out of the five runs we encountered an unrecoverable failure. In one of the runs, due to the instability of the grasping trajectory and the robot not tracking it perfectly, the fingers of the robot ended up pushing the milk away during grasping, which resulted in a very unstable grasp. As a result, the box fell to the ground in the carrying phase. Although during the table setting the robot was able to pick up a toppled over cup and successfully bring it to the table, picking up the milk box from the ground was impossible for the PR2. The other unrecoverable failure was the dishwasher grid getting stuck in PR2’s finger. Another major failure happened when placing the cereal box into its vertical drawer, which was difficult because the robot had to reach very high and approach its joint limits. When the gripper opened, the box fell on a side in the shelf, which resulted in it being crushed when the drawer was closed.
Photos: EASE
Failure cases including unstably grasping the milk, getting stuck in the dishwasher, and crushing the cereal.
While we’re focusing a little bit on the failures here, that’s really just to illustrate the exceptionally challenging edge cases that the robot encountered. Again, I want to emphasize that while the PR2 was not successful all the time, its performance over 90 minutes of fully autonomous operation is still very impressive. And I really appreciate that the researchers committed to an experiment like this, putting their robot into a practical(ish) environment doing practical(ish) tasks under full autonomy over a long(ish) period of time. We often see lots of incremental research headed in this general direction, but it’ll take a lot more work like we’re seeing here for robots to get real-world useful enough to reliably handle those critical breakfast tasks.
The Robot Household Marathon Experiment, by Gayane Kazhoyan, Simon Stelter, Franklin Kenghagho Kenfack, Sebastian Koralewski and Michael Beetz from the CRC EASE at the Institute for Artificial Intelligence in Germany, was presented at ICRA 2021. Continue reading
#437386 Scary A.I. more intelligent than you
GPT-3 (Generative Pre-trained Transformer 3), is an artificial intelligence language generator that uses deep learning to produce human-like output. The high quality of its text is very difficult to distinguish from a human’s. Many scientists, researchers and engineers (including Stephen … Continue reading
#439110 Robotic Exoskeletons Could One Day Walk ...
Engineers, using artificial intelligence and wearable cameras, now aim to help robotic exoskeletons walk by themselves.
Increasingly, researchers around the world are developing lower-body exoskeletons to help people walk. These are essentially walking robots users can strap to their legs to help them move.
One problem with such exoskeletons: They often depend on manual controls to switch from one mode of locomotion to another, such as from sitting to standing, or standing to walking, or walking on the ground to walking up or down stairs. Relying on joysticks or smartphone apps every time you want to switch the way you want to move can prove awkward and mentally taxing, says Brokoslaw Laschowski, a robotics researcher at the University of Waterloo in Canada.
Scientists are working on automated ways to help exoskeletons recognize when to switch locomotion modes — for instance, using sensors attached to legs that can detect bioelectric signals sent from your brain to your muscles telling them to move. However, this approach comes with a number of challenges, such as how how skin conductivity can change as a person’s skin gets sweatier or dries off.
Now several research groups are experimenting with a new approach: fitting exoskeleton users with wearable cameras to provide the machines with vision data that will let them operate autonomously. Artificial intelligence (AI) software can analyze this data to recognize stairs, doors, and other features of the surrounding environment and calculate how best to respond.
Laschowski leads the ExoNet project, the first open-source database of high-resolution wearable camera images of human locomotion scenarios. It holds more than 5.6 million images of indoor and outdoor real-world walking environments. The team used this data to train deep-learning algorithms; their convolutional neural networks can already automatically recognize different walking environments with 73 percent accuracy “despite the large variance in different surfaces and objects sensed by the wearable camera,” Laschowski notes.
According to Laschowski, a potential limitation of their work their reliance on conventional 2-D images, whereas depth cameras could also capture potentially useful distance data. He and his collaborators ultimately chose not to rely on depth cameras for a number of reasons, including the fact that the accuracy of depth measurements typically degrades in outdoor lighting and with increasing distance, he says.
In similar work, researchers in North Carolina had volunteers with cameras either mounted on their eyeglasses or strapped onto their knees walk through a variety of indoor and outdoor settings to capture the kind of image data exoskeletons might use to see the world around them. The aim? “To automate motion,” says Edgar Lobaton an electrical engineering researcher at North Carolina State University. He says they are focusing on how AI software might reduce uncertainty due to factors such as motion blur or overexposed images “to ensure safe operation. We want to ensure that we can really rely on the vision and AI portion before integrating it into the hardware.”
In the future, Laschowski and his colleagues will focus on improving the accuracy of their environmental analysis software with low computational and memory storage requirements, which are important for onboard, real-time operations on robotic exoskeletons. Lobaton and his team also seek to account for uncertainty introduced into their visual systems by movements .
Ultimately, the ExoNet researchers want to explore how AI software can transmit commands to exoskeletons so they can perform tasks such as climbing stairs or avoiding obstacles based on a system’s analysis of a user's current movements and the upcoming terrain. With autonomous cars as inspiration, they are seeking to develop autonomous exoskeletons that can handle the walking task without human input, Laschowski says.
However, Laschowski adds, “User safety is of the utmost importance, especially considering that we're working with individuals with mobility impairments,” resulting perhaps from advanced age or physical disabilities.
“The exoskeleton user will always have the ability to override the system should the classification algorithm or controller make a wrong decision.” Continue reading