Population-based learning techniques have been proven to be effective in dealing with noise and are thus promising tools for the optimization of robotic controllers, which have inherently noisy performance evaluations...
详细信息
ISBN:
(纸本)9781479914869
Population-based learning techniques have been proven to be effective in dealing with noise and are thus promising tools for the optimization of robotic controllers, which have inherently noisy performance evaluations. This article discusses how the results and guidelines derived from tests on benchmark functions can be extended to the fitness distributions encountered in robotic learning. We show that the large-amplitude noise found in robotic evaluations is disruptive to the initial phases of the learning process of PSO. Under these conditions, neither increasing the population size nor increasing the number of iterations are efficient strategies to improve the performance of the learning. We also show that PSO is more sensitive to good spurious evaluations of bad solutions than bad evaluations of good solutions, i.e., there is a non-symmetric effect of noise on the performance of the learning.
The ability to move in complex environments is a fundamental requirement for robots to be a part of our daily lives. Increasing the controller complexity may be a desirable choice in order to obtain an improved perfor...
详细信息
The ability to move in complex environments is a fundamental requirement for robots to be a part of our daily lives. Increasing the controller complexity may be a desirable choice in order to obtain an improved performance. However, these two aspects may pose a considerable challenge on the optimization of robotic controllers. In this paper, we study the trade-offs between the complexity of reactive controllers and the complexity of the environment in the optimization of multi-robot obstacle avoidance for resource-constrained platforms. The optimization is carried out in simulation using a distributed, noise-resistant implementation of Particle Swarm Optimization, and the resulting controllers are evaluated both in simulation and with real robots. We show that in a simple environment, linear controllers with only two parameters perform similarly to more complex non-linear controllers with up to twenty parameters, even though the latter ones require more evaluation time to be learned. In a more complicated environment, we show that there is an increase in performance when the controllers can differentiate between front and backwards sensors, but increasing further the number of sensors and adding non-linear activation functions provide no further benefit. In both environments, augmenting reactive control laws with simple memory capabilities causes the highest increase in performance. We also show that in the complex environment the performance measurements are noisier, the optimal parameter region is smaller, and more iterations are required for the optimization process to converge.
This paper addresses the problem of keeping an autonomous marine vehicle in a moving triangular formation by regulating its position with respect to two leader vehicles. The follower vehicle has no prior knowledge of ...
详细信息
The ability to move in complex environments is a fundamental requirement for robots to be a part of our daily lives. While in simple environments it is usually straightforward for human designers to foresee the differ...
详细信息
Self-assembly is a key coordination mechanism for large multi-unit systems and a powerful bottom-up technology for micro/nanofabrication. Controlled self-assembly and dynamic reconfiguration of large ensembles of micr...
详细信息
Self-assembly is a key coordination mechanism for large multi-unit systems and a powerful bottom-up technology for micro/nanofabrication. Controlled self-assembly and dynamic reconfiguration of large ensembles of microscopic particles can effectively bridge these domains to build innovative systems. In this perspective, we present SelfSys, a novel platform for the automated control of the fluidic self-assembly of microparticles. SelfSys centers around a water-filled microfluidic chamber whose agitation modes, induced by a coupled ultrasonic actuator, drive the assembly. Microparticle dynamics is imaged, tracked and analyzed in real-time by an integrated software framework, which in turn algorithmically controls the agitation modes of the microchamber. The closed control loop is fully automated and can direct the stochastic assembly of microparticle clusters of preset dimension. Control issues specific to SelfSys implementation are discussed, and its potential applications presented. The SelfSys platform embodies at microscale the automated self-assembly control paradigm we first demonstrated in an earlier platform.
This paper addresses the problem of keeping an autonomous marine vehicle in a moving triangular formation by regulating its position with respect to two leader vehicles. The follower vehicle has no prior knowledge of ...
详细信息
This paper addresses the problem of keeping an autonomous marine vehicle in a moving triangular formation by regulating its position with respect to two leader vehicles. The follower vehicle has no prior knowledge of the path described by the leaders but has access to their heading angle and is able to measure inter-vehicle ranges. It is assumed that the distance between the leaders is constant and known. A control strategy is adopted that generates speed and heading commands so as to drive suitably defined along track and cross track errors to zero. The commands are used as input to local inner loops for yaw and speed control. The paper describes the algorithms derived for range-based control and assesses their performance in simulations using realistic models of the vehicles involved. Tests with three autonomous marine vehicles equipped with acoustic modems and ranging devices allow for the evaluation of the performance of the algorithms in a real-world situation.
This paper presents a method for odor plume tracking by a swarm of robots in realistic conditions. In real world environments, the chemical concentration within an odor plume is patchy, intermittent and time-variant. ...
详细信息
This paper presents a method for odor plume tracking by a swarm of robots in realistic conditions. In real world environments, the chemical concentration within an odor plume is patchy, intermittent and time-variant. This study shows that swarm robots can cooperatively track the odor plume towards its source by establishing a cohesive spatial sensor network to deal with the turbulences and patchy nature of odor plumes. The robots move together and maintain a distance margin between themselves in order to keep the cohesion of the constructed sensor network while the odor concentration and air-flow speed are considered in the equations of navigation of the robots in the network to more efficiently track the plume. The method is evaluated in simulation against various number of robots, the emission rate of the odor source, the number of obstacles in the environment and the size of the testing environment. The emergent behavior of the swarm proves the functionality, robustness and scalability of the system in different conditions.
The design of high-performing robotic controllers constitutes an example of expensive optimization in uncertain environments due to the often large parameter space and noisy performance metrics. There are several eval...
详细信息
ISBN:
(纸本)9781479904532
The design of high-performing robotic controllers constitutes an example of expensive optimization in uncertain environments due to the often large parameter space and noisy performance metrics. There are several evaluative techniques that can be employed for on-line controller design. Adequate benchmarks help in the choice of the right algorithm in terms of final performance and evaluation time. In this paper, we use multi-robot obstacle avoidance as a benchmark to compare two different evaluative learning techniques: Particle Swarm Optimization and Q-learning. For Q-learning, we implement two different approaches: one with discrete states and discrete actions, and another one with discrete actions but a continuous state space. We show that continuous PSO has the highest fitness overall, and Q-learning with continuous states performs significantly better than Q-learning with discrete states. We also show that in the single robot case, PSO and Q-learning with discrete states require a similar amount of total learning time to converge, while the time required with Q-learning with continuous states is significantly larger. In the multi-robot case, both Q-learning approaches require a similar amount of time as in the single robot case, but the time required by PSO can be significantly reduced due to the distributed nature of the algorithm.
This work is developed in the framework of Institutional Robotics (IR), an approach to cooperative distributed robotic systems that draws inspiration from the social sciences. We consider a case study concerned with a...
详细信息
ISBN:
(纸本)9781467363563
This work is developed in the framework of Institutional Robotics (IR), an approach to cooperative distributed robotic systems that draws inspiration from the social sciences. We consider a case study concerned with a swarm of simple robots which has to maintain wireless connectivity and a certain degree of spatial compactness. Robots have local, bounded communication capabilities and have to execute the task (running an IR controller) using exclusively as information their current number of wireless connections to neighbors. For the very same case study, we previously introduced an IR-based macroscopic model for the behavior of a large number of robots, validated using a submicroscopic model implemented through a realistic simulator. In this work, we go a step further and validate our submicroscopic model with real world experiments, duplicating accurately the conditions used, including a large number of robots and noisy communication channels. The main conclusions of this paper are two-fold. First, the IR approach was able to maintain the wireless connectivity of a swarm of 40 real, resource-constrained robots. This speaks in favor of the robustness and scalability of such approach. Second, the submicroscopic model implemented is faithfully capturing the reality and can be used to further optimize the performances of distributed control strategies using an IR approach.
Environmental processes are often severely over-sampled. As sensor networks become more ubiquitous for this purpose, increasing network longevity becomes ever more important. Radio transceivers in particular are a gre...
详细信息
ISBN:
(纸本)9781479902064
Environmental processes are often severely over-sampled. As sensor networks become more ubiquitous for this purpose, increasing network longevity becomes ever more important. Radio transceivers in particular are a great source of energy consumption, and many networking algorithms have been proposed that seek to minimize their use. Traditionally, such approaches are often data agnostic, i.e., their performance is not dependent on the properties of the data they transport. In this paper we explore algorithms that exploit environmental relationships in order to reduce the amount of transmitted data while maintaining expected levels of accuracy. We employ a realistic testing environment for evaluating the power savings brought by such algorithms, based on Sensorscope, a commercial sensor network product for environmental monitoring. We implement and test a suppression-based data collection algorithm from literature that to our knowledge has never been implemented on a real system, and propose modifications that make it more suitable for real-world conditions. Using a custom extension board developed for in situ power monitoring, we show that while the algorithms greatly reduce the amount of energy spent on transmitting packets, they have no effect on the real system's overall power consumption due to its preexisting network architecture.
暂无评论