A key challenge in evolving control systems for robots using neural networks is training tractability. Evolving monolithic fixed topology neural networks is shown to be intractable with limited supervision in high dimensional search spaces. Common strategies to overcome this limitation are to provide more supervision by encouraging particular solution strategies, manually decomposing the task and segmenting the search space and network. These strategies require a supervisor with domain knowledge and may not be feasible for difficult tasks where novel concepts are required. The alternate strategy is to use self-organized task decomposition to solve difficult tasks with limited supervision. The artificial neural tissue (ANT) approach presented here uses self-organized task decomposition to solve tasks. ANT inspired by neurobiology combines standard neural networks with a novel wireless signaling scheme modeling chemical diffusion of neurotransmitters. These chemicals are used to dynamically activate and inhibit wired network of neurons using a coarse-coding framework. Using only a global fitness function that does not encourage a predefined solution, modular networks of neurons are shown to self-organize and perform task decomposition. This approach solves the sign-following task found to be intractable with conventional fixed and variable topology networks. In this paper, key attributes of the ANT architecture that perform self-organized task decomposition are shown. The architecture is robust and scalable to number of neurons, synaptic connections, and initialization parameters.
Important tasks such as environment monitoring require field devices such as sensors that can operate for long durations. Current power supply technologies such as batteries limit many applications. Fuel cells are a promising alternative to batteries because they can have much higher energy densities. However, their lives may be short due to catalyst degradation. Here, a simplified model of proton exchange membrane (PEM) fuel cell catalyst degradation is applied to small fuel cells. The model focuses on the combined effects of catalyst dissolution and migration. The effect of migration on catalyst degradation is found to be substantial and this has not been accounted for in previous models. The model considers the effect of field conditions such as varying power demands, temperature and humidity, and predicts the catalyst life of the fuel cell and its power output. The predicted life is a proposed metric that can quantify the relative importance and effect of field conditions on the catalyst particularly for the design and control of fuel cell power supplies. Experiments are presented that support the model. This model is applied to a study on field sensors and results suggests unless PEM fuel cells are isolated from damaging field conditions, they will have short lives.
Abstract-Autonomous control systems onboard planetary rovers and spacecraft benefit from having cognitive capabilities like learning so that they can adapt to unexpected situations in-situ. Q-learning is a form of reinforcement learning and it has been efficient in solving certain class of learning problems. However, embedded systems onboard planetary rovers and spacecraft rarely implement learning algorithms due to the constraints faced in the field, like processing power, chip size, convergence rate and costs due to the need for radiation hardening. These challenges present a compelling need for a portable, low-power, area efficient hardware accelerator to make learning algorithms practical onboard space hardware. This paper presents a FPGA implementation of Q-learning with Artificial Neural Networks (ANN). This method matches the massive parallelism inherent in neural network software with the fine-grain parallelism of an FPGA hardware thereby dramatically reducing processing time. Mars Science Laboratory currently uses Xilinx-Space-grade Virtex FPGA devices for image processing, pyrotechnic operation control and obstacle avoidance. We simulate and program our architecture on a Xilinx Virtex 7 FPGA. The architectural implementation for a single neuron Q-learning and a more complex Multilayer Perception (MLP) Q-learning accelerator has been demonstrated. The results show up to a 43-fold speed up by Virtex 7 FPGAs compared to a conventional Intel i5 2.3 GHz CPU. Finally, we simulate the proposed architecture using the Symphony simulator and compiler from Xilinx, and evaluate the performance and power consumption.
Autonomous robotic excavation has often been limited to a single robotic platform using a specified excavation vehicle. This paper presents a novel method for developing scalable controllers for use in multirobot scenarios and that do not require human defined operations scripts nor extensive modeling of the kinematics and dynamics of the excavation vehicles. Furthermore, the control system does not require specifying an excavation vehicle type such as a bulldozer, frontloader or bucket-wheel and it can evolve to select for an appropriate choice of excavation vehicles to successfully complete a task. The "Artificial Neural Tissue" (ANT) architecture is used as a control system for autonomous multirobot excavation and clearing tasks. This control architecture combines a variable-topology neural-network structure with a coarsecoding strategy that permits specialized areas to develop in the tissue. Training is done in a low-fidelity grid world simulation environment and where a single global fitness function and a set of allowable basis behaviors need be specified. This approach is found to provide improved training performance over fixedtopology neural networks and can be easily ported onto different robot platforms. Aspects of the controller functionality have been tested using high fidelity dynamics simulation and in hardware. An evolutionary training process discovers novel decentralized methods of cooperation employing aggregation behaviors (via synchronized movements). These aggregation behaviors are found to improve controller scalability (with increasing robot density) and better handle robot interference (antagonism) that reduces the overall efficiency of the group.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.