Robotic Systems Integration for Supply Chain Workflows: Design, Deploy, Execute

Many emerging robotics applications require the use of multiple collaborating robots to operate under human supervision. To be useful in such applications, collaborative robots will need to:

• program themselves
• learn efficiently from the observed performance
• operate safely in the presence of uncertainty
• call for help appropriately during the execution of challenging tasks
• communicate effectively with humans

This presentation will provide an overview of the advances in physics-aware artificial intelligence that enable robots to automatically make decisions to meet these requirements, including:

• an approach for automatically generating near-optimal trajectories in real time to enable robots to program themselves from task descriptions
• self-directed learning methods to equip robots with the ability to learn from observing the performance of previously executed tasks and adapting their plans
• methods for robots to operate safely in the presence of uncertainty by generating contingency-aware plans
• computational methods for endowing robots with introspective capabilities so that they can seek help from humans on challenging tasks
• augmented reality-based interfaces for enabling robots to elicit human guidance during the decision-making process

The use of collaborative robots can significantly improve human productivity and eliminate the need for human involvement in tasks that pose risks to human safety. Assembly, composite layup, machine tending, and sanding tasks will be used as illustrative examples to show how smart collaborative robots can be used in high-mix manufacturing applications.

Developing industrial robotics requires knowledge and experience in many engineering domains, including mechanical design, perception, decision making, control design, and embedded systems. In this talk, I will discuss how we can design, test, validate and verify, and deploy these complex industrial robotics applications. Several concrete examples, from pick-and-place manipulation to warehouse robot coordination, illustrate how the complexity of these systems can be managed through scalable physics simulation, Model-Based Design, and hardware connectivity and deployment.

Experiential robotics is aimed at enhancing human experiences in all aspects of life from work to home to common places. This presentation will introduce the principles of experiential robotics for the factory of the future with use cases from seafood processing and personal protective equipment manufacturing.

In this talk, I will introduce our recent projects of Model-Based Design and control of soft robots for practical applications.

Current trends show a rise in the use of automated robots in manufacturing industries such as aerospace, automotive, biomedical, and clean energy. Multi-agent systems are a key concept in the larger megatrend of robotics in warehouse automation. In this talk, we will explore how to design, simulate, and scale multi-agent systems by evaluating objectives such as sizing a workspace, optimizing available resources, and maximizing work.

In this talk, we will present a mixed-reality interface for multi-human multi-robot interaction. Multi-robot systems are complex systems composed of many interacting entities. These systems exceed the typical human attention span, which studies place at seven (plus or minus two) entities in laboratory conditions. A natural approach to improving human performance is to relieve the burden of individual operators by conceiving supervisory control schemes in which multiple humans cooperate. However, with multiple humans in the system, additional challenges arise including unbalanced workload, inhomogeneous awareness, and conflict among operators. This limits the performance of the operators, typically measured in terms of workload, situational awareness, and trust in the system. We will show that the performance of human users can be improved with a mixed-reality interface that offers mixed granularity of control, increased transparency, and mixed communication modalities among operators. The interface was validated experimentally throughout eight user studies involving a total of 122 participants, in which we analyzed operator workload, awareness, trust, and performance.

A predictive approach to maintenance operations guarantees advantages such as reducing machine downtime, eliminating unnecessary interventions, and adding revenue streams for equipment vendors with aftermarket services. These benefits are achievable, but there are challenges in implementing predictive maintenance in application fields such as industrial robotics.

One obstacle is the limited availability of functioning data and the lack of failure data needed for accurate results. In addition, choosing the best models for monitoring and prediction can be time consuming, and integrating devised algorithms with existing tools and infrastructures can cause a bottleneck in the implementation process.

In this talk, we will analyze the benefits and challenges of predictive maintenance, and describe, via the exploration of a real case-study, the solution proposed by MathWorks to overcome these obstacles and realize efficient predictive maintenance systems.

The digitalization of manufacturing is enabling the modularization and flexibilization of production processes in Industry 4.0 by enhancing the convergence of the digital and physical worlds. It is also helping to manage the growing complexity of the production facilities introduced by global competition and sophisticated products. Robotic manipulation systems are used in production processes to increase the level of automation. By adopting model-based systems engineering (MBSE), the functional correctness and safety concerns of a given design of robotic manipulation systems can be validated against the system requirements as digital assets before the actual physical deployment in production facilities. The resulting engineering life cycle can then be significantly reduced due to the front loading of the system validation and effective use of the digital assets.

This talk will review recent progress in vision-based grasping—a common task in many processes and one that is performed effortlessly by humans. Early robotic approaches were heavy on planning and required the world to be static or moving in a simple and predictable way. It is now possible to close the perception-action loop in milliseconds without requiring explicit prior models of the objects to be grasped. The talk will touch on our entry in the Amazon Robotics Challenge, recent work in learning-based grasping, and current work on visual-loop closing of finger-tip motion and applications to food factories of the future.