A mechatronics approach to copier paperpath design

Today's copier machines are optimized for a narrow range of sheet properties, as specified by the manufacturer. As a result, they often perform poorly when using media with non-standard characteristics, such as recycled paper. This follows from the fact that copier machines are predominantly controlled in an open loop fashion, i.e. there is little or no error correction during sheet transport. When a machine detects that a sheet has deviated from its nominal position, it shuts down and signals the user to remove all sheets and start over again. This lack of reliability manifests itself more often with faster machines.; This dissertation addresses the aforementioned robustness problem with a mecha-tronics design paradigm. By modifying the copier paperpath layout, adding extra sensors and computer control, it becomes possible to perform closed loop control of sheet positions.; The dissertation presents two paperpath layout redesigns and corresponding control strategies. The tight connection between paperpath layout, controller design complexity and achievable performance, guides the discussion. Only longitudinal sheet position errors are considered.; A first proposed paperpath layout consists of multiple independent rollers. Sheets can be individually controlled, which simplifies controller design. Sheet position errors are removed using polynomial reference trajectories. In addition, a design algorithm minimizes the required paperpath length, subject to constraints on the trajectories.; A second proposed paperpath layout reduces the number of required actuators by grouping rollers into independent sections. Controller design is now complicated by the hybrid dynamics of the paperpath. Sheets residing in the same section are no longer independently controllable.; Dynamic programming fits the problem definition very well. It provides a centralized optimal control strategy in table look-up form. A corresponding reachability analysis provides further insight into layout design issues. The implementation is limited to simplified dynamics. Excessive computational and memory requirements prevent extension of the results to more complete dynamics and therefore limit its practical use.; A different control approach is then developed using mere heuristics. The result is a distributed, hybrid hierarchical control strategy which controls the spacings between sheets.; The heuristic control strategy is implemented on an experimental setup, developed and built at the University of California at Berkeley. The setup uses standard copier parts, arranged as three independently actuated sections. An observer fuses the measurements from multiple optical sensors into sheet position estimates for closed loop control. The control software is developed using graphical tools with automatic code generation in a rapid protoyping environment.; The results show robust performance and precision for a wide variety of media at state-of-the-art throughputs.