based on principles from modern artificial intelligence and robotics, we developed the RoboMusic concept. In RoboMusic, we use a number of robotic devices as instruments, and the tunes are composed as a behavior-based...
详细信息
based on principles from modern artificial intelligence and robotics, we developed the RoboMusic concept. In RoboMusic, we use a number of robotic devices as instruments, and the tunes are composed as a behavior-based system. The music artist composes a baseline behavior for the robotic instruments, and composes the behavioral response to interactions by human musicians. The music artist is transformed from a composer of static music tunes to a developer of robot behavior: behavior that is expressed by the robotic system as music pieces. Music compositions are transformed to become robotic behavior as in a behavior-based system. A RoboMusic concert is performed with robotic instruments, and changes the concept of live concerts by inviting the audience to interact with the band's instruments themselves and thereby guide the live performance of the music themselves.
Most of the proposed approaches for behavior coordination neglect the issue of learning from experience, which is essential to make an agent capable of adapting to the environment. This subject is mostly addressed by ...
详细信息
Most of the proposed approaches for behavior coordination neglect the issue of learning from experience, which is essential to make an agent capable of adapting to the environment. This subject is mostly addressed by researchers in the area of Reinforcement Learning. In order to investigate the capability of Reinforcement Learning techniques for behavior arbitration, we considered a sub-problem in robotic soccer. We have solved this problem by using Sarsa(lambda) algorithm with a pseudo-fuzzy technique for function approximation. We also compared Sarsa(lambda) with Q(lambda), and CMAC function approximation with the proposed APF technique.
We propose a modular architecture for autonomous robots which allows for the implementation of basic behavioral modules by both programming and training, and accommodates for an evolutionary development of the interco...
详细信息
We propose a modular architecture for autonomous robots which allows for the implementation of basic behavioral modules by both programming and training, and accommodates for an evolutionary development of the interconnections among modules. This architecture can implement highly complex controllers and allows for incremental shaping of the robot behavior. Our proposal is exemplified and evaluated experimentally through a number of mobile robotic tasks involving exploration, battery recharging and object manipulation.
It is important to methodologically distinguish between the environment dynamics, the agent dynamics and their coupling. In order to make this distinction clearer, let us define the following vocabulary. We will call ...
详细信息
It is important to methodologically distinguish between the environment dynamics, the agent dynamics and their coupling. In order to make this distinction clearer, let us define the following vocabulary. We will call a sensory-motor loop the internal mechanism linking perception to command whether fixed or adaptive and however sophisticated the percepts and commands can be for the sake of generality. We will call behavior the externally observed behavior tin the intuitive sense) produced by the execution of such a sensory-motor loop in a particular context. The originality of this work is, first, the formulation of reinforcement learning based on the sensory-motor dynamics and not the environment dynamics, and, second, to present a flexible selection learning mechanism of several sensory-motor loops and its formulation as a case of adaptive optimal control. Selection is presented as a way to coordinate independently developed sensory-motor loops.
Maintaining contact between the robot and plume is significant in chemical plume tracing (CPT). In the time immediately following the loss of chemical detection during the process of CPT, Track-Out activities bias the...
详细信息
Maintaining contact between the robot and plume is significant in chemical plume tracing (CPT). In the time immediately following the loss of chemical detection during the process of CPT, Track-Out activities bias the robot heading relative to the upwind direction, expecting to rapidly re-contact the plume. To determine the bias angle used in the Track-Out activity, we propose an online instance-based reinforcement learning method, namely virtual trail following (VTF). In VTF, action-value is generalized from recently stored instances of successful Track-Out activities. We also propose a collaborative VTF (cVTF) method, in which multiple robots store their own instances, and learn from the stored instances, in the same database. The proposed VTF and cVTF methods are compared with biased upwind surge (BUS) method, in which all Track-Out activities utilize an offline optimized universal bias angle, in an indoor environment with three different airflow fields. With respect to our experimental conditions, VTF and cVTF show stronger adaptability to different airflow environments than BUS, and furthermore, cVTF yields higher success rates and time-efficiencies than VTF.
One of the core challenges of long-term autonomy is the environmental dynamics that agents must interact with. Some of these dynamics are driven by reliable cyclic processes, and thus are predictable. The most dominan...
详细信息
One of the core challenges of long-term autonomy is the environmental dynamics that agents must interact with. Some of these dynamics are driven by reliable cyclic processes, and thus are predictable. The most dominant of these is the daily solar cycle, which drives both natural phenomena like weather, as well as the activity of animals and humans. Circadian clocks are a widespread solution in nature to help organisms adapt to these dynamics, and demonstrate that many organisms benefit from maintaining simple models of their environments and how they change. Drawing inspiration from circadian systems, this work models relevant environmental states as times series, allowing for forecasts of the state to be generated without any knowledge of the underlying physics. These forecasts are treated as special percepts in a behavior-based architecture;providing estimates of the future state rather than measurements of the current state. They are incorporated into an ethologically based action-selection mechanism, where they influence the activation levels of behaviors. The approach was validated on a simulated agricultural task: a solar-powered agent monitoring pest populations. By using the artificial circadian system to leverage the forecasted state, the agent was able to improve performance and energy management.
based on the concepts of RoboMusic and modular playware, we developed a system composed of modular playware devices which allow any user to perform music in a simple, interactive manner. The key features exploited in ...
详细信息
based on the concepts of RoboMusic and modular playware, we developed a system composed of modular playware devices which allow any user to perform music in a simple, interactive manner. The key features exploited in the modular playware approach are modularity, flexibility, construction, immediate feedback to stimulate engagement, creative exploration of play activities, and in some cases activity design by end-users (e.g., DJs). We exemplify the approach with the development of 11 rock genres and 6 pop music pieces for modular I-BLOCKS, which are exhibited and in daily use at the Rock Me exhibition, and have been used at several international music events in Japan and the USA. A key finding is that professional music design is essential for the development of primitives in a musical behavior-based system, and this professional esthetics is necessary to engage the users in the activity of assembling and coordinating these "professional" musical primitives. This article describes, explores, and discusses this concept.
Deliberate control of an entertainment robot presents a special problem in balancing the requirement for intentional behavior with the existing mechanisms for autonomous action selection. It is proposed that the inten...
详细信息
Deliberate control of an entertainment robot presents a special problem in balancing the requirement for intentional behavior with the existing mechanisms for autonomous action selection. It is proposed that the intentional biasing of activation in lower-level reactive behaviors is the proper mechanism for realizing such deliberative action. In addition, it is suggested that directed intentional bias can result in goal-oriented behavior without subsuming the underlying action selection used to generate natural behavior. This objective is realized through a structure called the intentional bus. The intentional bus serves as the interface between deliberative and reactive control by realizing high-level goals through the modulation of intentional signals sent to the reactive layer. A deliberative architecture that uses the intentional bus to realize planned behavior is described. In addition, it is shown how the intentional bus framework can be expanded to support the serialization of planned behavior by shifting from direct intentional influence for plan execution to attentional triggering of a learned action sequence. Finally, an implementation of this architecture, developed and tested on Sony's humanoid robot QRIO, is described.
From an early stage in their development, human infants show a profound drive to explore the objects around them. Research in psychology has shown that this exploration is fundamental for learning the names of objects...
详细信息
From an early stage in their development, human infants show a profound drive to explore the objects around them. Research in psychology has shown that this exploration is fundamental for learning the names of objects and object categories. To address this problem in robotics, this paper presents a behavior-grounded approach that enables a robot to recognize the semantic labels of objects using its own behavioral interaction with them. To test this method, our robot interacted with 100 different objects grouped according to 20 different object categories. The robot performed 10 different behaviors on them, while using three sensory modalities (vision, proprioception and audio) to detect any perceptual changes. The results show that the robot was able to use multiple sensorimotor contexts in order to recognize a large number of object categories. Furthermore, the category recognition model presented in this paper was able to identify sensorimotor contexts that can be used to detect specific categories. Most importantly, the robot's model was able to reduce exploration time by half by dynamically selecting which exploratory behavior should be applied next when classifying a novel object. (C) 2012 Elsevier B.V. All rights reserved.
暂无评论