Navigation of autonomous mobile robot using different activation functions of wavelet neural network

Panigrahi, Pratap Kumar; Ghosh, Sandip; Parhi, Dayal R.

doi:10.1515/acsc-2015-0002

Cited by 9 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This procedure is repeated until the last layer is reached, where the signal W L x L−1 + B L is sent as an output directly without applying the activation function (see Figure 1). There are many commonly used activation functions, see, e.g., [1,16,25,28,29] for further discussion.…”

Section: Feedforward Neural Network (Fnns)mentioning

confidence: 99%

Approximation properties of Residual Neural Networks for Kolmogorov PDEs

Jonas¹,

Salimova²

2021

Preprint

View full text Add to dashboard Cite

In recent years residual neural networks (ResNets) as introduced by He et al. [17] have become very popular in a large number of applications, including in image classification and segmentation. They provide a new perspective in training very deep neural networks without suffering the vanishing gradient problem. In this article we show that ResNets are able to approximate solutions of Kolmogorov partial differential equations (PDEs) with constant diffusion and possibly nonlinear drift coefficients without suffering the curse of dimensionality, which is to say the number of parameters of the approximating ResNets grows at most polynomially in the reciprocal of the approximation accuracy ε > 0 and the dimension of the considered PDE d ∈ N. We adapt a proof in Jentzen et al. [20] who showed a similar result for feedforward neural networks (FNNs) -to ResNets. In contrast to FNNs, the Euler-Maruyama approximation structure of ResNets simplifies the construction of the approximating ResNets substantially. Moreover, contrary to [20], in our proof using ResNets does not require the existence of an FNN (or a ResNet) representing the identity map, which enlarges the set of applicable activation functions.

show abstract

Section: Feedforward Neural Network (Fnns)mentioning

confidence: 99%

Approximation properties of Residual Neural Networks for Kolmogorov PDEs

Jonas¹,

Salimova²

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The paper 79 has discussed about path planning using adaptive neuro-fuzzy inference system for mobile robot. Wavelet neural network for navigation of mobile robot has been discussed by Panigrahi et al [80][81][82] have focused on adaptive neuro fuzzy technique for control of mobile robot. Pham et al 83 have used neural network technique for path planning of mobile robots.…”

Section: Analysis Of Various Ai Techniques Used For Navigationmentioning

confidence: 99%

Advancement in navigational path planning of robots using various artificial and computing techniques

Parhi¹

2018

IRATJ

View full text Add to dashboard Cite

In this paper, a review analysis has been carried out on navigational methodologies of robots with the help of various artificial intelligence techniques such as Fuzzy Logic, Neural Network, Genetic Algorithm, Particle Swarm Optimisation (PSO) and other Artificial Intelligence techniques. During the survey analysis a systematic review has been done and shown how various artificial intelligence techniques can be successfully used for control and navigation of various types of robots subjected to various environmental conditions.

show abstract

“…Panigrahi et al (2015) proposed a neural network (NN)-based path planning for a mobile robot. This method is based on neurons and the network present in the human brain.…”

Section: Introductionmentioning

confidence: 99%

Optimal path planning of mobile robot using the hybrid cuckoo–bat algorithm in assorted environment

Murali

Deepak

M.B.L.

et al. 2019

IJIUS

View full text Add to dashboard Cite

Purpose The purpose of this paper is to obtain an optimal mobile robot path planning by the hybrid algorithm, which is developed by two nature inspired meta-heuristic algorithms, namely, cuckoo-search and bat algorithm (BA) in an unknown or partially known environment. The cuckoo-search algorithm is based on the parasitic behavior of the cuckoo, and the BA is based on the echolocation behavior of the bats. Design/methodology/approach The developed algorithm starts by sensing the obstacles in the environment using ultrasonic sensor. If there are any obstacles in the path, the authors apply the developed algorithm to find the optimal path otherwise reach the target point directly through diagonal distance. Findings The developed algorithm is implemented in MATLAB for the simulation to test the efficiency of the algorithm for different environments. The same path is considered to implement the experiment in the real-world environment. The ARDUINO microcontroller along with the ultrasonic sensor is considered to obtain the path length and time of travel of the robot to reach the goal point. Originality/value In this paper, a new hybrid algorithm has been developed to find the optimal path of the mobile robot using cuckoo search and BAs. The developed algorithm is tested with the real-world environment using the mobile robot.

show abstract

Navigation of autonomous mobile robot using different activation functions of wavelet neural network

Cited by 9 publications

References 8 publications

Approximation properties of Residual Neural Networks for Kolmogorov PDEs

Approximation properties of Residual Neural Networks for Kolmogorov PDEs

Advancement in navigational path planning of robots using various artificial and computing techniques

Optimal path planning of mobile robot using the hybrid cuckoo–bat algorithm in assorted environment

Contact Info

Product

Resources

About