Simulating Robots Without Conventional Physics: A Neural Network Approach

Pretorius, Christiaan J.; Plessis, Mathys C. du; Cilliers, Charmain

doi:10.1007/s10846-012-9782-6

Cited by 25 publications

(9 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SNNs have been successfully utilised in the development of controllers for simple differentially steered mobile robots. 4,5,18 Problems solved include path planning, obstacle avoidance and lightapproaching behaviours. The use of SNNs for simulating the behaviours of an inverted pendulum robot has been successfully demonstrated.…”

Section: Simulator Neural Networkmentioning

confidence: 99%

“…The use of SNNs for simulating the behaviours of an inverted pendulum robot has been successfully demonstrated. 4,16 SNNs can serve as a fast surrogate model to a more high-fidelity simulation where evaluations can be time-consuming. 19 Prior investigations into the use of SNNs for simulating the behaviours of a complex snake-like robot required structured changes to the joint angles.…”

Section: Simulator Neural Networkmentioning

confidence: 99%

“…1,2 The ER process has been successfully applied to many problems, such as obstacle avoidance, light following, trajectory planning, inverted pendulum stabilisation, evolutionary-added design and many others. [3][4][5][6] A large number of controller evaluations are required during the ER process in order to rank relative fitnesses of controllers. Evaluating a large number of controllers on a real-world robot is often time-consuming and may damage hardware through mechanical wear.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Complex Morphology Neural Network Simulation in Evolutionary Robotics

Woodford

Plessis

2019

Robotica

View full text Add to dashboard Cite

SUMMARYThis paper investigates artificial neural network (ANN)-based simulators as an alternative to physics-based approaches for evolving controllers in simulation for a complex snake-like robot. Prior research has been limited to robots or controllers that are relatively simple. Benchmarks are performed in order to identify effective simulator topologies. Additionally, various controller evolution strategies are proposed, investigated and compared. Using ANN-based simulators for controller fitness estimation during controller evolution is demonstrated to be a viable approach for the high-dimensional problem specified in this work.

show abstract

Section: Simulator Neural Networkmentioning

confidence: 99%

Section: Simulator Neural Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Complex Morphology Neural Network Simulation in Evolutionary Robotics

Woodford

Plessis

2019

Robotica

View full text Add to dashboard Cite

show abstract

“…The most affected field is probably evolutionary robotics because of the emphasis on opening the search space as much as possible: behavior found within the simulation is often not anticipated by the designer of the simulator, therefore, it's not surprising that they are often wrongly simulated. Researchers in evolutionary robotics explored three main ideas to cross this 'reality gap': (1) automatically improving simulators (Bongard et al, 2006;Klaus et al, 2012;Pretorius et al, 2012), (2) trying to prevent optimized controllers from relying on the unreliable parts of the simulation (in particular, by adding noise) (Jakobi et al, 1995), and (3) modeling the difference between simulation and reality (Hartland and Bredeche, 2006;Koos et al, 2012).…”

Section: Concept and Intuitionsmentioning

confidence: 99%

Fast damage recovery in robotics with the T-resilience algorithm

Koos

Cully

Mouret

2013

The International Journal of Robotics Research

View full text Add to dashboard Cite

Damage recovery is critical for autonomous robots that need to operate for a long time without assistance. Most current methods are complex and costly because they require anticipating potential damage in order to have a contingency plan ready. As an alternative, we introduce the T-resilience algorithm, a new algorithm that allows robots to quickly and autonomously discover compensatory behavior in unanticipated situations. This algorithm equips the robot with a self-model and discovers new behavior by learning to avoid those that perform differently in the self-model and in reality. Our algorithm thus does not identify the damaged parts but it implicitly searches for efficient behavior that does not use them. We evaluate the T-resilience algorithm on a hexapod robot that needs to adapt to leg removal, broken legs and motor failures; we compare it to stochastic local search, policy gradient and the self-modeling algorithm proposed by Bongard et al. The behavior of the robot is assessed on-board thanks to an RGB-D sensor and a SLAM algorithm. Using only 25 tests on the robot and an overall running time of 20 min, T-resilience consistently leads to substantially better results than the other approaches

show abstract

“…Thanks to its ability to make predictions without a full mapping of the mechanistic details, deep learning has been used to emulate time-consuming model simulations [28][29][30][31] . To date, however, the predicted outputs are restricted in categorical labels or a set of discrete values.…”

Section: Introductionmentioning

confidence: 99%

Massive computational acceleration by using neural networks to emulate mechanism-based biological models

Wang

Fan

Luo

et al. 2019

Preprint

View full text Add to dashboard Cite

Mechanism-based mathematical models are the foundation for diverse applications. It is often critical to explore the massive parametric space for each model. However, for many applications, such as agent-based models, partial differential equations, and stochastic differential equations, this practice can impose a prohibitive computational demand. To overcome this limitation, we present a fundamentally new framework to improve computational efficiency by orders of magnitude. The key concept is to train an artificial neural network using a limited number of simulations generated by a mechanistic model. This number is small enough such that the simulations can be completed in a short time frame but large enough to enable reliable training of the neural network. The trained neural network can then be used to explore the system dynamics of a much larger parametric space. We demonstrate this notion by training neural networks to predict self-organized pattern formation and stochastic gene expression. With this framework, we can predict not only the 1-D distribution in space (for partial differential equation models) and probability density function (for stochastic differential equation models) of variables of interest with high accuracy, but also novel system dynamics not present in the training sets. We further demonstrate that using an ensemble of neural networks enables the self-contained evaluation of the quality of each prediction. Our work can potentially be a platform for faster parametric space screening of biological models with user defined objectives.

show abstract

Simulating Robots Without Conventional Physics: A Neural Network Approach

Cited by 25 publications

References 28 publications

Complex Morphology Neural Network Simulation in Evolutionary Robotics

Complex Morphology Neural Network Simulation in Evolutionary Robotics

Fast damage recovery in robotics with the T-resilience algorithm

Massive computational acceleration by using neural networks to emulate mechanism-based biological models

Contact Info

Product

Resources

About