Photo credit: www.therobotreport.com
ABBs is working with T-Robotics and Mbodi to enhance the autonomy and intuitive capabilities of robots. Source: ABB Robotics
The application of artificial intelligence in industrial robotics holds the promise of creating machines that are not only smarter but also easier to operate. However, as innovators aspire to reach this goal, they often face challenges. Recently, ABB Robotics announced that Mbodi and T-Robotics have emerged as the winners of its 2024 ABB Robotics AI Startup Challenge. Out of over 100 global entrants, the two startups were selected from a competitive pool, where 15 were chosen to pitch their ideas and five submitted minimum viable product (MVP) proposals.
“Modern manufacturing demands solutions that blend user-friendliness with advanced AI capabilities,” said Marc Segura, president of ABB Robotics. “The innovations presented by T-Robotics and Mbodi are truly pioneering, and our collaboration with them aims to fast-track the development of robotic solutions that are more intuitive, flexible, and efficient, ultimately enhancing business resilience.”
Segura elaborated that the challenge focused on exploring the potential of natural language interfaces and enabling robots to autonomously plan tasks. “We’re considering broader applications rather than just specific technical KPIs for dexterous tasks or navigation. We aim to engage startups that are at a high technology readiness level [TRL] and poised for market entry.”
As part of their win, both startups will be awarded $30,000 in project funding and will work closely with ABB’s technology experts. In addition, each company will benefit from a six-month membership to SynerLeap, ABB’s dedicated startup accelerator program.
“Understanding our customers’ challenges regarding performance, safety, and return on investment is crucial,” Segura noted. “The surge of generative AI requires numerous iterations, and we’re keen to partner with startups that possess core competencies to morph these innovations into commercially viable products.”
Both T-Robotics and Mbodi.ai are aiming to roll out their initial commercial applications in collaboration with ABB in the United States by 2025.
Register today to save 40% on conference passes!
T-Robotics: Advancing Robotic Communication
Traditionally, tasks such as programming a CNC machine were accomplished through cumbersome methods, but T-Robotics aims to modernize that process. “With our technology, users can visually teach the robot tasks, like directing it to pick an item from a tote, and even instruct it to open a door if necessary,” Segura explained.
Utilizing physical AI models, T-Robotics has created a natural language interface tailored for robotics. The San Francisco and Trondheim-based startup asserts that this method can drastically cut down programming time while optimizing performance across diverse manufacturing applications.
“Our system’s ability to understand semantics through generative AI, followed by a deterministic decision tree approach, enhances precision, which is pivotal for safety,” Segura added. “We’re exploring complementary methodologies for this purpose.”
Asad Tirmizi, founder of T-Robotics, expressed optimism about the partnership with ABB, stating, “The fusion of robotics with physical AI is set to revolutionize manufacturing processes, and we are excited to usher in an era where robots serve as intelligent assistants, enhancing production capabilities.”
Source: ABB Robotics
Mbodi: Enabling Task Execution
The focus of Mbodi’s AI platform lies in the effective execution of tasks. “We employ transformer algorithms and an agentic infrastructure that empowers robots to learn and adapt to new tasks almost instantaneously through both written directives and live demonstrations,” Segura pointed out.
Operating from New York, Mbodi emphasizes that its technology is a significant leap toward democratizing automation for companies, especially those engaged in high-mix, low-volume production scenarios.
“We’re enriching the industrial robotics landscape with advanced generative AI that responds to the evolving demands of customers,” stated Sebastian Peralta, co-founder and CEO of Mbodi. “ABB’s commitment to groundbreaking technologies like ours is vital to fostering this transformation.”
Segura clarified that they are not overly ambitious. “We’re aiming for practical applications that enable targeted automation in specific domains. For instance, instead of training a neural network for an entire-assignment like electronics assembly, we intend to empower manufacturers to instruct robots in specific tasks such as wiring adjustments.”
Accuracy in robot training also involves combining visual data with tactile feedback for comprehensive skill training, Segura noted, emphasizing a more granular approach rather than creating an all-encompassing robotic solution.
Mbodi anticipates that vision language action models (VLAs) could enable robots to generalize tasks when integrated with current technologies like machine vision and cloud computing.
Source: ABB Robotics
ABB’s Commitment to Practical AI Solutions
ABB is deeply invested in fostering startup engagement and collaboration through ABB Robotics & Automation Ventures (ABB RA Ventures), an initiative designed to leverage industry expertise to innovate and invest in early-stage companies with global significance.
“The potential of generative AI in the robotics sector is substantial, but it necessitates careful examination,” Segura noted. “By centering on key areas like natural language programming, skill acquisition, and autonomous decision-making, we have spotlighted two main solutions that will expedite our development of market-ready, AI-driven industrial products.”
The ABB Robotics AI Startup Challenge allowed participants the freedom to explore various models, although MVPs could utilize ABB’s Robot Studio on OmniCore for enhanced cloud-based control and collaboration, as explained by Segura.
This challenge is part of ABB’s broader Innovation Ecosystem, which collaborates with both established and emerging businesses to accelerate innovation and launch new products that contribute to a sustainable and productive future.
“Our framework encompasses stakeholders, startups, and specific challenges for discourse,” continued Segura. “ABB actively engages with diverse companies that bring unique application expertise, such as Covariant for picking solutions and Sevensense for visual SLAM, the latter of which ABB has acquired.”
Segura concluded, “Our customers are actively seeking ways to reduce the costs and complexities associated with automation. Generative AI is key to empowering robots to operate more effectively while decreasing operational costs, and these objectives are interconnected.”
Source
www.therobotreport.com