Pick-and-place machines are essential tools in modern automation, used in everything from electronics manufacturing to packaging. However, many current systems lack “precise generalization” – the ability to handle a variety of tasks without sacrificing accuracy. This often leads to highly specialized, inflexible solutions.
“In industry, you often see that [manufacturers] end up with very tailored solutions to the particular problem that they have, so a lot of engineering and not so much flexibility in terms of the solution,” explains Maria Bauza Villalonga, a senior research scientist at Google DeepMind.
SimPLE solves this problem and provides a solution to pick-and-place that is flexible and still provides the needed precision.
A new study published in Science Robotics by researchers from MIT’s Department of Mechanical Engineering introduces SimPLE (Simulation to Pick Localize and placE), a novel approach to pick-and-place that addresses this limitation. SimPLE allows robots to accurately pick, regrasp, and place objects using their CAD models, even without prior experience with those specific objects.
Alberto Rodriguez, an MIT visiting scientist and associate director of manipulation research for Boston Dynamics, highlights the key advantage of SimPLE: “The promise of SimPLE is that we can solve many different tasks with the same hardware and software using simulation to learn models that adapt to each specific task.”
SimPLE achieves its precision through a combination of task-aware grasping, visuotactile perception (using both sight and touch), and regrasp planning. The system utilizes a dual-arm robot equipped with visuotactile sensors. Real-world observations are compared to simulated data through supervised learning, allowing the robot to estimate object poses and execute precise placements.
In experimental testing, SimPLE successfully picked and placed a variety of objects with different shapes, achieving success rates exceeding 90% for 6 objects and 80% for 11 objects. Antonia Delores Bronars, a mechanical engineering doctoral student, emphasizes the significance of incorporating tactile sensing: “There’s an intuitive understanding in the robotics community that vision and touch are both useful, but [until now] there haven’t been many systematic demonstrations of how it can be useful for complex robotics tasks.”
Experts in the field have recognized the value of SimPLE’s approach. Matt Mason, chief scientist at Berkshire Grey and professor emeritus at Carnegie Mellon University, notes, “Most work on grasping ignores the downstream tasks. This paper goes beyond the desire to mimic humans, and shows from a strictly functional viewpoint the utility of combining tactile sensing, and vision, with two hands.”
Ken Goldberg, a distinguished chair in engineering at UC Berkeley, sees SimPLE as a valuable alternative to the current trend of relying solely on AI and machine learning: “The authors combine well-founded geometric algorithms that can reliably achieve high-precision for a specific set of object shapes and demonstrate that this combination can significantly improve performance over AI methods. This can be immediately useful in industry and is an excellent example of what I call ‘good old fashioned engineering’ (GOFE).”
The development of SimPLE highlights the power of collaboration and building upon previous research. As Bronars points out, “Collaboration, with each other and with Nikhil Chavan-Dafle and Yifan Hou, and across many generations and labs really allowed us to build an end-to-end system.”
Responses (0 )