Methodology

 

Methodology

Though a structured and coherent approach, ROSSINI will develop and demonstrate technologies enabling a significant advancement in HRC. Besides implementing 4 synergic lines of research, and the integration of the results into one comprehensive platform for the design and validation of HRC application, ROSSINI will also develop 3 industrial demonstrators, which will act as technological showcases for the market replication and therefore for the full leverage of the market potential of exploitable results. These are the 4 distinct lines of research of the Rossini Project:

 
 
 

Sensing Technology

The RS4 (ROSSINI Smart and Safe Sensing System) will be the link between the environment and an efficient and safe movement of a robot in it, when at least a human operator is also present in the same environment.


Robot Control Technology

In ROSSINI safety is transformed from an unforeseen barrier into a dynamic constraint to consider when dynamically planning the best sequence of actions to fulfil a desired task.


Robot Actuation Technology

The ultimate goal of the project is to combine the best of both worlds to engineer a new generation of robots featuring the safety peculiarities of cobots and the high performances of industrial robots.


Human-robot mutual understanding

Successful adoption of new and genuine human-robot collaborations, both increasing flexible production and improving the quality of the job

 
 
 

Sensing Technology

RS4 will allow both a 2D and a 3D monitoring based on the specific requirements. 2D and 3D sensors can be used and each sensor may have a different performance in terms of speed, resolution, accuracy. In order to avoid occlusions or to improve safety some areas might be covered by redundant sensors.

RS4 will first of all include a new Safe 3D Vision Sensor module, for which Hardware, Software, Mechanical and Optical with intrinsic safety features will be carried out. The PILZ Safety Eye represents the State of the Art and it will be taken as the starting point for a design that aims to be a substantial improvement. The whole sensors array will need standard ethernet based safety fieldbus communication, with modular hardware implementation development using standard FPGAs or Microprocessors. Finally RS4 will involve the design of a Safety Sensor modules Controller, in charge of the integration of multi sensors information in a single multidimensional image of the overall scene (fusion of RS4 sensors partial data).

The vision is to implement speed and distance monitoring exploiting different technologies, in order to develop a system compliant to ISO/TS 15066:2106 standard.

 
 
 
 
 
 
 

Robot Control Technology

ROSSINI aims at radically changing the current safety paradigm adopted in the setup of cooperative robotic cells for industrial application. Currently, safety is treated as a barrier in front of which the robot has to slow down and eventually stop. The behaviour (e.g. motion, interaction) of the robot is the output of a planner whose planning strategy is built considering a free environment and/or few obstacles in a fixed position. Because of the variability of the environment, such a setup can cause several unexpected stops and the user has to adapt to the safety-induced, often inefficient for a specific cooperative scenario, behaviour of the robot during the cooperation. In ROSSINI safety is transformed from an unforeseen barrier into a dynamic constraint to consider when dynamically planning the best sequence of actions to fulfil a desired task. This makes ROSSINI collaborative robotic system much more flexible and efficient than collaborative systems currently available on the market. ROSSINI will reverse the classical paradigm according to which “the user has to learn how the robot works and to adapt to it” for achieving a novel cooperative paradigm according to which “the robot learns what the human wants and adapts to it”.

To this aim, the data coming from the sensors of the cooperative cells will be collected and aggregated to achieve a semantic scene map that allows the control system to be aware of the position of the main elements of the cooperative task to execute (e.g. humans, objects to manipulate) and the main areas where a safe behaviour is required (e.g. humans, infrastructure elements, mobile robots). Semantic scene maps are dynamically updated and explicitly considered in the design of the ROSSINI controllers in order to build a safety aware control architecture. By the knowledge of the task to execute, of the input of the human and of the safety critical areas, the control architecture can dynamically optimize the behaviour of the robot for maximizing the efficiency while preserving the safety of the overall system.

 
 
 
 

Robot Actuation Technology

The ultimate goal of the project is to combine the best of both worlds to engineer a new generation of robots featuring the safety peculiarities of cobots and the high performances of industrial robots. Specifically, the focus must be placed on the ability of handling precisely and fast heavy objects with a collaborative robot and concurrently guarantee a high safety level for the human operator. In order to achieve the above mentioned goal, the following aspects have to be considered in the project:

 
3.png

a. Robot mechanical conceptual redesign.

A major mechanical redesign is needed in order to reduce the overall inertia of the robotic arm and ensure a stiffness level which eventually results in a positioning precision comparable to the one of standard industrial robots.

b. Force feedback

In most cases force detection is used to make the robot more collaborative. If an external force is applied to the robot, (by a human operator, for example), it is detected through motor current allowing the controller to make a decision on an alternative response like stop motion, reduce speed, or change direction. The new approach uses dual encoders, not only to improve joint accuracy, but for real time monitoring of the stiffness/compliance in each robot joint. Through this method, joint position and torque can be monitored together to provide safe information to the collaborative robot controller. With accurate torque and position monitoring force sensing, even in the presence of stiffness/compliance can be compensated for.

c. Dual-motor robot joint

In order to increase the intrinsic safety of the collaborative robot, the basic idea is to design a new concept of robot joint, provided with two motors, the first one responsible of the normal joint positioning during robot task execution, the second one acting as safety device for fast-retract of the robot in case of collision with a human.

 
 
 

With the newly designed robot 3 factors contribute to the reduction of the stopping distance:

  • dual-motor joint with fast braking and retraction (20% expected reduction)

  • dual-encoder joint with force feedback fast evaluation (10% expected reduction)

  • mechanical robot redesign with reduction of total robot inertia (15% expected reduction)

 
 
 

Human-robot mutual understanding

Successful adoption of new and genuine human-robot collaborations, both increasing flexible production and improving the quality of the job, requires a holistic approach in which:

  1. human factors like user experience, comfort, trust, feeling of safety, and liability, are addressed and accounted for in the early design stages (Design Level);

  2. constant monitoring of the process, human behaviour, and robot behaviour takes place and online changes to the original task planning can be made during operation (Adaptive Level);

  3. profound mutual understanding between robots and people in operation is realized (Communication Level).

 
collab hands 2-01.png

Design level

Elements to address in the design stage, is the explicit design of the human-robot interaction and the definition of human-robot collaboration scenarios. Here the work process is evaluated. Based on a task and capacity analysis it is investigated which actors (human and/or robot) can perform which task. Moreover, scenarios describe possible ways for humans and robots to interact. The scenarios indicate what information needs to be exchanged between actors to establish mutual understanding and successful job completion (Johnson et al., 2014). Furthermore the scenarios give a basic indication on how the workspace is designed (shared workspace, synchronous movements, basic safety implementations). An evaluation tool allows to assess job quality, productivity, flexibility, and configuration time for different collaboration scenarios in early stages of the design process.

 
 

Adaptive level

The adaptive level dispatches tasks to the actors according to the scenarios that were made. When the scenarios contain multiple execution paths the ACL should consider human and machine factors when dispatching tasks to humans or robots. Dispatching criteria are influenced by foreseeable and unforeseeable factors.

  • Foreseeable: Inclusiveness of vulnerable workers, Day shifts / night shifts, training, scheduled equipment maintenance, etc.

  • Unforeseeable: Sick leave, safety interventions, equipment failure, temporarily lowering of operator capacity, etc.

Communication level

To enhance smooth human robot interaction, human and robot must be mutually predictable and adequately estimate each other’s intentions (Klein et al, 2004). Two key technologies to achieve this are:

  • Estimating human intentions through sensor fusion

  • Projecting robot intentions through augmented reality (AR)