A heuristically derived stabilization strategy for an unstable and unintuitive plant by fuzzy control is described. It is shown that the often used classical fuzzy controller, which is both static and time invariant, is incapable of stabilizing such types of plants. However, a simple modification to the classical fuzzy controller architecture that separates the measurement and control phases, together with a hierarchical control strategy, enable the unstable and unintuitive plant to be stabilized. The fuzzy control strategy, as well as the new fuzzy controller architecture, are based on the consideration of "what a human subject would do when dealing with a physical plant which is both unstable and unintuitive". The stabilization strategy is then generalized to other mathematically similar systems. While the rules for the stabilization of the plant are heuristically defined, the membership functions associated with the rules are tuned by a simulated annealing procedure.

The authors outline the development of a robust controller for the cooperative multiarm manipulation of a single object. Based on a derived model of closed-chain dynamic behavior, a controller consisting of nonlinear compensation terms together with an acceleration feedback structure is proposed. This controller has the property that it simultaneously regulates the motion response and the internal force response of the object, for a given class of reference signals, over a wide range of parameter variations of the system. The desired speed of response is obtained by online tuning of two controller parameters. The effectiveness of the proposed control method was demonstrated by a simulation of two three-link planar arms jointly manipulating a rigid bar.>

An experimental study of a central- izad/decentralized controller for rigid robotic manipu- lators, which solves the robust servomechanism prob- lem for reference signals which asymptotically become constant, is presented. The controller has two tun- parameters which are tuned, using on-line tun- ing methods, so as to optimize the system's speed of response. Exact knowledge of the manipulator dy- namics is not required to implement the controller; in particular, no detailed knowledge of the payload mass is required, i.e. the controller is highly robust to parameter variations of the robotic system. Accelera- tion feedback is the key feature which produces such robust behaviour in the controller. The proposed controller is implemented on a pla- nar five-bar parallel linkage robot and compared to an independent joint PID controller.

Some comments on a recent published paper dealing with structural controllability are made.

An arbitrary interconnected network of intersecting streets with one-way traffic flow is considered in which some or all intersections are assumed to be operating at maximum capacity. In this case, it is desired to find a robust decentralized controller at each traffic intersection, so that for all perturbed external traffic flow sources/sinks, the resulting perturbed queues at each intersection are asymptotically “balanced”. The only local knowledge required of the controllers at each intersection is that re the queue lengths of the intersection. It is shown that the necessary and sufficient conditions for a solution to the problem to exist depend solely on the topology of the network and the percentage of vehicles turning at each intersection, and are identical to the conditions required for a solution to exist for the centralized controller case. In particular, it is shown that if turns always occur at an intersection, then a solution to the problem always exists. A characterization of controllers which solve the problem are then given. In particular, it is shown that the local controllers which solve the problem are identical to each other, which implies that if the network is expanded (assuming a solution exists) then one can use the same local controllers to solve the problem. It is also shown that the local controllers can, in principle, give “perfect control” to the problem. Some examples are included to illustrate the results.

A generalized Crank–Nicholson technique is given for the solution of simultaneous first-order linear time-invariant differential equations. The method proposed is inherently stable for any time interval h, gives correct solutions for any h as t → ∞ for polynomial type time inputs and has a truncation error of O(h5). It is especially suitable for the solution of large systems of differential equations.