Home Blog Page 6

Multitasking Dung Beetle Inspired Robots

0

Plenty of researchers  are inspired by nature when improving the design of their own robots. Binggwong Leung, Sanislav Gorb, and Poramate Manoonpong have published a paper that explores dung beetle-like robot (ALPHA) with biomechanical feet that can perform adaptive walking and ball rolling tasks on flat and uneven terrains. It can handle different ball types and heights.

Nature's All-in-One: Multitasking Robots Inspired by Dung Beetles

As the researchers explain, the modular modular neural-based “contains a central pattern generator (CPG) module, a pattern formation network (PFN) module, and a robot orientation control (ROC) module”

[HT]

UBHOME M10 Robotic Mower with LoRa, RTK&AI Vision

0

This is the UBHOME M10: a smart robotic mower with RTK, AI Vision and LoRa connectivity that can cover up to 7000m2 of lawn every week. It doesn’t require a perimeter wire to work. The M10 has 9 sensors to detect obstacles and get around them. It has an adjustable cutting height of 1.18 to 2.36″. You can use your phone to remote control the robot and manage multiple zones.

SUNSEEKER Orion X7+ AI Vision Robotic Mower with Smart Path Planning

0

This is the SUNSEEKER X7+: a smart robotic mower that you can put to use without using any boundary wires. It uses RTK-GNSS and VSLAM to get around on your property while avoiding obstacles. It has a dual blades for precise cuts. The Orion X7 Plus can cover 1.5 acre of land. It has a cutting width of 14 inches and cutting height of 0.8 to 4 inches.

Robots Can Now Seal Doors To Prevent Hazardous Material Leakage

0

In the past few years, we have covered plenty of robots that can put out fires while keeping humans safe. They can also be used to seal a fire door with a sealant. This is useful in nuclear facilities where leaks of hazardous materials can pose serious danger to field operators.

RoadRunner M&M: Learning Based Framework for Autonomous Robot Navigation

0

Autonomous off-road navigation can be challenging for robots. It requires understanding of conditions and geometric information. RoadRunner M&M is an end-to-end learning-based framework that predicts “the traversability and elevation maps at multiple ranges.” As input, it takes a voxel map and multiple images. Robots using this framework will be able to assess their surroundings in high detail.

RoadRunner M&M: Dense Forest Environment (Out-Of-Distribution)

According to the researchers, this approach achieves up to 50% better elevation mapping and 30% enhanced traversability estimation over the previous version.

[HT]

LucidSim: Robot Learning Visual Parkour from Generated Images

0

Here is a framework to train quadruped robots using only RGB camera inputs. LucidSim creates diverse and physically accurate image sequences from the robot’s perspective. Training is done within simulated environments. Generative models are used to generate realistic visual data, enhancing the robot’s ability to generalize across different environments.

Ropet: AI Robot Pet with Lifelike Movement

0

Meet the Ropet robot pet: an AI robot that can dance to your music and engage you in chats thanks to ChatGPT. It has a soft touch fur and maintains a body temperature similar to humans. This gorgeous robot has expressive eyes. It is smart enough to recognize over 100s of foods. It develops a unique personality as you interact with it.

Ropet Al Robot Pet (Coming Soon on Kickstarter)

As you interact with this robot, you earn points to unlock eye and voice effects. You can also customize its fur color. This robot has a high definition camera, stepper motors, speaker, and touch sensors. You can find this project on Kickstarter.

Unitree G1 Open Source Dataset Released

0

The Unitree G1 is one of the most impressive humanoid robots we have seen. It is very agile and can perform long jumps up to 1.4 meters. It uses a 3D LiDAR and depth camera to perceive its surroundings. Its operation data set is now open sourced. That includes learning algorithms, datasets, and models.

DEEPRobotics Lynx Extreme Off-road Robot

0

In the past few years, we have covered plenty of robots that can adapt to off-road conditions. The DEEPRobotics Lynx robot is one of the most impressive ones yet. This wheeled robot dog can do backflips, run up and down challenging terrains and even maintain its balance in bipedal mode. The below video shows it in action:

Extreme Off-Road | DEEPRobotics Lynx All-Terrian Robot

[HT]

RoboPAIR: jailbreaking LLM-Driven Robots

0

In the past few months, we have seen plenty of robots that offer integration with large language models. While LLMs revolutionize robots with contextual reasoning and facilitate human-robot interaction, they open robots to risk of being jailbroken. RoboPAIR is an algorithm designed to jailbreak LLM-controlled robots. It “elicits harmful physical actions from LLM-controlled robots.” Here is what the researchers accomplished:

  • White-box setting: Full access to NVIDIA Dolphins self-driving LLM.
  • Gray-box setting: Partial access to Clearpath Robotics Jackal UGV with GPT-4o planner.
  • Black-box setting: Query access to GPT-3.5-integrated Unitree Robotics Go2.

According to the researchers, in many scenarios, they managed to achieve 100% attack success rate. In the above video, you can see how the robot was tricked to deliver an explosive package.

[HT]

More Robots: