8.04 Digital technologies for human centric and collaborative human-machine manufacturing systems

REFERENCE SPOKE
OTHER SPOKES
PROJECT LEADER
Paolo Rocco
START DATE
Gennaio 2023
END DATE
Dicembre 25
PROPOSER
Politecnico di Milano
PARTNERS

Università degli Studi di Padova, Sapienza Università di Roma, Camozzi Group S.p.A. 

8.04 Digital technologies for human centric and collaborative human-machine manufacturing systems

Agile manufacturing is a relatively new term adopted to describe a production approach able to respond quickly to unforeseen customer demands, market volatilities, or other factors of high manufacturing impact such as changing lot sizes, variants, process technologies. The project concerns the design of monitoring, control and supervision techniques for the optimal management of time and resources in the digital factory. Particular emphasis will be given to the use of (mobile-based) collaborative robots and to the flexible use of agents (machineries, robots, operators, etc.). Among the various possible applications of collaborative robotics, the assembly of mechanical parts is of particular importance.

The project will be organized along three main lines of research:

  1. A first area is related to the pre-deployment of the application. Robotic applications like an assembly one have to be accurately designed and engineered in order to take optimal advantages from both the human and the robot. Part presentation, robot placement, robot movements, human and robot tasks represent typical decision variables. On the other hand, human safety and ergonomics and cycle times are typical KPIs. The design task of application engineers may be assisted by edge/cloud automated algorithms (capable of performing several what-if analyses in background) suggesting the engineering team suitable modifications of the layout/application for a more productive application.
  2. A second area concerns the execution of the digital factory at a macro level. We are making reference here to the whole orchestration of the process, handled through a high-level control loop. The orchestration instructions concern the optimal dispatching and scheduling of the single activities among the different agents (humans and robots). Such instructions will be generated by a centralized cloud-based algorithm, through the evaluation of relevant data obtained at the shopfloor and then shared.
  3. A third area concerns again the execution of the digital factory, however now at a micro level. Reference is then made to the single collaborative cell where. The availability of agents’ digital twins will allow the possibility to monitor the workers in the shopfloor and then help in the early discover of ergonomic issues. On the other hand, the composition of agents’ digital twins at factory-level will help in deriving optimal manufacturing operation management policies (dispatching, scheduling, etc.) to support production agility. The micro level will also be pushed at the level of the single component. Development of intelligent and adaptive components that gather data at edge or local level and understand whether or not a certain operation is correctly executed will be addressed as well.

The data-driven nature of all the three aspects in the design of the digital factory exposes the system to security, privacy, and potentially safety concerns. Evaluation and design of a threat and adversary model, alongside data protection and cybersecurity measures, with special focus on the cyber-physical elements of the system, will be developed during the project.

RISULTATI ATTESI

The final aim of this project is to build data-driven tools, possibly based on cloud-based artificial intelligence, for control, resource allocation, and scheduling in order to make a production line agile, flexible, easily reconfigurable, and secure.

Specifically, the following results are foreseen in the 36 months’ time-frame of the project:

  1. Layout design for human-robot shared working environment with task scheduling and balancing (both micro and macro layout, both manufacturing and intralogistics areas)
  2. Development of edge/cloud automated algorithms that can suggest an application engineer the adaptations and modifications of the layout of a robotized assembly line for a more productive application
  3. Development of factory-level digital twins to be used in optimal manufacturing operation management policies (dispatching, scheduling, etc.) to support production agility.
  4. Study and development of an orchestrator of the digital factory (mixed human-robot production line) in the form of a centralized cloud-based algorithm. The orchestrator is data driven, in the sense that it evaluates relevant data obtained at the shopfloor with the help of the digital twin, and produces optimal scheduling and dispatching policies.
  5. Methods and tools for anticipation of future worker motion intention in order to optimize the cobot’s motion planning and task execution when human and cobots are collaborating
  6. Development of agents’ digital twins to monitor the workers in the shopfloor for the assessment of ergonomic issues. Digital ergo assessment of human-robot collaboration in manufacturing and logistics by using digital ergonomics tools for postural real time assessment and fatigue and recovery models validated by EMG technology.
  7. Methods for the estimation of the interaction forces in cobotic applications
  8. Methodologies to make the component, and in particular the robot end-effector, intelligent, with relocation of some of the digital factory functionalities at edge level or at component level.
  9. Evaluation and design of a threat and adversary model with special focus on the cyber-physical elements of the system.