This paper presents a new self-learning control scheme for lateral track guidance of industrial trucks using artificial intelligence. It is an universally applicable lateral dynamic control concept which is able to adapt itself to different truck variants. Moreover it shall consider vehicle parameter variations that occur during operation, such as the load dependent change of vehicle mass and moment of inertia. The proposed approach uses Reinforcement Learning. In order to reduce the training effort, a new concept is realized, taking into account a priori knowledge of vehicle behavior. Its fundamental idea consists of dividing the training process into two steps. In the first step the controller will be pre-trained on basis of a nominal model representing a priori knowledge of lateral dynamic vehicle behavior. Since this model is derived for an industrial truck with average vehicle parameter values, a fine tuning of the control parameters has to be performed in the second step. In this way the controller is adapted to the actual truck variant and the corresponding vehicle parameter values. In order to demonstrate the efficiency of the proposed control scheme, the simulation results given in this paper are compared to the closed loop behavior using standard LQR.
This paper presents an extension of a self-learning control concept for automatic track guidance of industrial trucks in intralogistic systems. The presented approach is based on Reinforcement Learning (RL), a method of Artificial Intelligence (AI) and is able to adapt itself to different
industrial truck variants and the associated specific vehicle parameters. Moreover, time-variant parameters during operation, such as the vehicle's velocity are taken into account. In order to consider the existing a priori knowledge of the controlled system and to avoid starting
the whole training process of the controller for each truck variant from scratch, the training process is divided into two steps. In the first step, the controller is trained on a model using
parameters of a nominal vehicle variant. Based on this, the control parameters are only finetuned in the second step. In this way the controller is adapted to the actual truck variant and the corresponding parameter values. In order to take into account the time-variant vehicle parameters during operation, the Artificial Neural Networks (ANN) of the RL controller and the observation vector are suitably extended. In this way, the varying speed can be considered in both training steps and the control parameters can be optimized accordingly. Thus, in case of
the investigated scenarios a stable control loop behavior can be guaranteed for the entire speed range of industrial trucks. In order to demonstrate this, the new approach is compared with a
RL control concept, not considering time-variant parameters.
This paper presents a new control concept for automatic track guidance of industrial trucks in intralogistic systems. It is based on Reinforcement Learning (RL), a method of Artificial Intelligence (AI). The presented approach is able to adapt itself to different industrial truck variants and to the associated specific vehicle parameters. In order to avoid starting the whole training of the controller for each truck variant from scratch, the training process is divided into two steps. In the first step, the controller is trained on a model using parameters of a nominal vehicle variant. Based on this, the control parameters are only fine-tuned in the second step. In this way the controller is adapted to the actual truck variant and the corresponding parameter values. Moreover, the influence of the disturbance variable of the system (path curvature) is compensated by considering this a priori knowledge within the control design. Therefore, the Artificial Neural Networks (ANN) of the RL controller and the observation vector are suitably adjusted. In this way, the occurring path curvatures can be considered in both training steps and the control parameters can be optimized accordingly. Thus, the influence of the disturbance variable can be compensated, which significanlty improves the control quality. In order to demonstrate this, the new approach is compared to a RL control concept, not considering the disturbance variable and to a classical two-degrees-of-freedom (2DoF) control approach.
This paper presents an extended control concept for automatic track guidance of industrial trucks in intralogistic systems. It is based on Reinforcement Learning (RL), a method of Artificial Intelligence (AI). The presented approach is able to adapt itself to different industrial truck variants and to the associated specific vehicle parameters. In order to avoid starting the whole
training of the controller for each truck variant from scratch, the training process is divided into two steps. In the first step, the controller is trained on a simplified linear model using parameters
of a nominal vehicle variant. Based on this, the control parameters are only fine-tuned in the second step using a more complex nonlinear model, representing the real industrial truck. In this
way, the controller is adapted to the actual truck variant and the corresponding parameter values. By using the nonlinear model, it can be ensured that the forklift’s dynamic is approximated
within the entire operating range, even at high steering angles. Moreover, the influence of the disturbance variable of the system (path curvature) is compensated by considering this a priori
knowledge within the control design. Therefore, the
Artificial Neural Networks (ANN) of the RL controller and the observation vector are suitably adjusted. In this way, the occurring path curvatures can be considered in both training steps and the control parameters can be optimized accordingly. Thus, the influence of the disturbance variable can be compensated, which significantly improves the control quality. In order to demonstrate this, the new approach is compared
to an RL control concept, which is not considering the disturbance variable and to a classical
two-degrees-of-freedom (2DoF) control approach.
Im Verbundprojekt KAnIS forscht die TH Aschaffenburg mit Linde Material Handling an neuen KI-basierten Algorithmen zur querdynamischen Fahrzeugführung von Flurförderzeugen. Ziel des Projekts ist die Entwicklung eines universell einsetzbaren Regelungskonzepts, das verschiedene Arten von Flurförderzeugen automatisch führt, im Betrieb auftretenden Fahrzeugparameterschwankungen gezielt Rechnung trägt und darüber hinaus auch den jeweiligen Fitnesszustand der Fahrzeuge berücksichtigt.
In Zeiten einer globalen Weltwirtschaft und einem von zunehmendem Wettbewerb geprägten Markt ist die Automatisierung logistischer Prozesse eine Voraussetzung für den Unternehmenserfolg. Vor allem die Steigerung der Produktivität und der Effizienz des innerbetrieblichen Materialflusses nimmt einen hohen Stellenwert ein. Mit dem Einsatz automatisch fahrender Flurförderzeuge (FFZ) lässt sich dies erreichen. Als Vision sollten jedoch nicht einzelne automatisch fahrende FFZ gelten. Ziel muss es sein, die Mitglieder einer heterogenen Logistikflotte miteinander zu vernetzen und eine Kommunikations- und Rechenplattform einzurichten. Dies bietet ein großes Potenzial im Hinblick auf die Auftragsplanung zur Verbesserung der Wirtschaftlichkeit sowie in den Bereichen Lokalisierung und Sicherheit zur Optimierung automatisch fahrender FFZ.
In this paper, a new control concept for automatic track guidance of forklifts within a heterogeneous logistics fleet is presented. The proposed control scheme is universally applicable and based on Artificial Intelligence (AI). The lateral control is realized using a Model Predictive Controller (MPC). In order to take into account the diversity of the industrial truck variants, a multi-model approach is used. Therefore, a plant model for each truck variant is integrated into the MPC. In a practical application, the most suitable model has to be selected. This decision is based on AI in the form of an Artificial Neural Network (ANN). It is able to distinguish the different truck variants based on that part of the plant's state vector, which represents the vehicle dynamic characteristics. Thus, the appropriate model can be used, which significantly improves the control quality and guarantees an accurate track guidance of different forklifts. Due to the AI-based selection of the model, the computational effort can be kept low and real-time capability can be ensured. In order to classify the performance of the proposed control concept, its simulation results are compared to the closed-loop behavior, using a classical MPC.
Two different control concepts for the automatic track guidance of forklifts are
proposed. Both approaches are based on Reinforcement Learning (RL), a method of Artificial
Intelligence (AI), and are able to take into account time-variant parameters, such as the vehicle velocity, and to reduce the influence of the path curvature, the most important disturbance variable of lateral vehicle control. In the first approach, both, the path curvature and the vehicle velocity signal, are provided to the controller in addition to the state variables of the controlled system. By varying the corresponding parameters in the training process, both signals can be considered and the control parameters can be optimized accordingly. In the second approach, several controllers (multi-model concept) considering the path curvature are used and the varying vehicle velocity is taken into account using a gain-scheduling concept. Considering time-variant vehicle parameters and the influence of the disturbance variable during operation, a stable track guidance is guaranteed within the whole speed range of the industrial trucks.