Flexible multi-sensorial system for automatic disassembly using cooperative robots

Abstract

Flexible multisensorial systems are a very important issue in the current industry when disassembling and recycling tasks have to be performed. These tasks can be performed by a human operator or by a robot system. In the current paper a robotic system to perform the required tasks is presented. This system takes into consideration the distribution of the necessary tasks to perform the disassembly of a component using several robots in a parallel or in a cooperative way. The algorithm proposed to distribute the task among robots takes into consideration the characteristics of each task and the sequence that needs to be followed to perform the required disassembly of the product. Furthermore, this paper presents a disassembly system based on a sensorized cooperative robots interaction framework for the planning of movements and detections of objects in the disassembly tasks. To determine the sequence of the disassembly of some products, a new strategy to distribute a set of tasks among robots is presented. Subsequently, the visual detection system used for detecting targets and characteristics is described. To carry out this detection process, different well known strategies, such as matching templates, polygonal approach and edge detection, are applied. Finally, a visual-force control system has been implemented in order to track disassembly trajectories. An important aspect of this system is the processing of the sensorial information in order to guarantee coherence. This aspect allows the application of both sensors, visual and force sensors, co-ordinately to disassembly tasks. The proposed system is validated by experiments using several types of components such as the covers of batteries and electronic circuits from toys, and drives and screws from PCs.

Disassembly Automation

“…Lambert [5] reviews typical AI techniques for disassembly sequence planning, including simulated annealing algorithms, genetic algorithms (GA), fuzzy sets, neural networks, multi-agent systems, and Bayesian networks. Other novel algorithms that have been efficiently applied to DSP include ant-colony optimisation [23], case-based reasoning [24] and rule-based sequence generation on clustering graphs [25].…”

Basic behaviour control of the vision‐based cognitive robotic disassembly automation

“…Regarding the perception carried out as part of the automation, the condition of the product is perceived throughout the disassembly process by the sensing modules, namely the vision system which is mainly used in many research works such as Torres et al (2004Torres et al ( , 2009, Gil et al (2007), Knoth et al (2002), Merdan et al (2010) and Berger and Schmidt (1995). The module provides information on the physical properties of the component, e.g.…”

A cooperative robotic system based on multiple sensors to construct metallic structures

“…During the assembly, the system must determine which robot performs the main task and which robot helps in its execution. The tasks to be developed by each robot are distributed using a task planner [6,9] which determines the actions for the assembly process. The robot R 1 has a parallel gripper, a screwdriver and a vacuum available and the robot R 2 has a Barret-hand and a range camera available.…”

“…The sensors which to detect if unexpected events appear during the development of the tasks, such as the presence of a human who crosses the trajectory planned for a robot, are also important. In previous works, some basic techniques of visual servoing and visual detection of target using image processing [6] have been applied to monitorize the movement of a robot and guide it to achieve the suitable position in the disassembly and manipulation processes. In addition, nowadays, human-robot interaction is used more and more to help in the modelling and localization of objects [7], and control techniques are applied more to manipulate and insert objects [8] in assembly processes.…”