The Robotics Roundup is a weekly newspost going over some of the most exciting developments in robotics over the past week.
In today’s edition we have:
- Giving Robots the Ability to Reason
- Disney showcasing latest robots at Robotics Summit & Expo
- Building trust between humans and robots when managing conflicting objectives
- Researchers help robots navigate efficiently in uncertain environments
- Mercedes is trialing humanoid robots for ‘low skill, repetitive’ tasks
Giving Robots the Ability to Reason
Covariant, an AI robotics company collaborating with retailers and logistics providers, has unveiled its new robotics foundation model, RFM-1. This model aims to equip robots with human-like reasoning abilities and a comprehensive understanding of language and the physical world. Covariant operates across multiple sectors, including apparel, health and beauty, and pharmaceuticals. The company’s co-founder and CEO, Peter Chen, discussed these developments on “Bloomberg Technology.”
Disney showcasing latest robots at Robotics Summit & Expo
Disney’s Imagineering team, known for bringing Disney characters to life through advanced robotics, will be featured at the Robotics Summit & Expo at the Boston Convention Center on May 1-2. Researchers Moritz Baecher, Morgan Pope, and Tony Dohi will showcase Disney’s latest robots and discuss the creation of robotic versions of renowned Disney characters. Each of these researchers brings a wealth of knowledge and experience in robotics, and their work focuses on designing and controlling robotic characters with believability and dynamic performances.
Building trust between humans and robots when managing conflicting objectives
A study from the University of Michigan has shown that trust and team performance in tasks with conflicting objectives improve when robots adapt to human strategies. This is the first such study to demonstrate these results. The researchers developed an algorithm for this purpose which can apply to any human-robot interaction scenario involving trade-offs, such as balancing a patient’s pain tolerance with their long-term health goals in a rehabilitation setting. The research aims to shift robots from being mere tools for isolated tasks to trustworthy collaborative partners. This approach was tested in a game-like scenario where a team comprising a human and robot had to balance completing a task quickly with maintaining health levels.
Researchers help robots navigate efficiently in uncertain environments
Researchers at MIT have developed an algorithm to help robots navigate efficiently in uncertain environments. The algorithm creates roadmaps of uncertain environments that strike a balance between quality and efficiency, enabling the robot to find a navigable route that minimizes travel time. It works by initially identifying safe paths and then discovering shortcuts to reduce overall travel time. This algorithm could be useful in exploration scenarios or in search-and-rescue missions. The researchers’ approach represents the environment as a graph and uses the Canadian Traveler’s Problem to minimize travel time with a given probability of success. This research, which outperformed baseline methods in over 100 simulated experiments, could have real-world applications in urban environments and complex robotic manipulation problems.
Mercedes is trialing humanoid robots for ‘low skill, repetitive’ tasks
Mercedes-Benz has partnered with Apptronik to trial the use of the company’s 160-pound bipedal robot, Apollo, in automating “low skill, physically challenging, manual labor” tasks within its vehicle production lines. Apptronik suggests that humanoid robots like Apollo could automate certain tasks without requiring a redesign of existing facilities, focusing on tasks for which it’s difficult to find reliable workers. The trials are taking place at a Mercedes factory in Hungary, a country that has faced labor shortages for several years. The move follows similar initiatives by other automakers like BMW and Tesla.