Path optimization for navigation of a humanoid robot using hybridized fuzzy-genetic algorithm

Asita Kumar Rath (Centre of Biomechanical Science, Institute of Technical Education and Research, Siksha O Anusandhan University, Bhubaneswar, India)
Dayal R. Parhi (Department of Mechanical Engineering, National Institute of Technology Rourkela, Rourkela, India)
Harish Chandra Das (Department of Mechanical Engineering, National Institute of Technology Meghalaya, Shillong, India)
Priyadarshi Biplab Kumar (Department of Mechanical Engineering, National Institute of Technology Rourkela, Rourkela, India)
Manoj Kumar Muni (Department of Mechanical Engineering, National Institute of Technology Rourkela, Rourkela, India)
Kitty Salony (Department of Electrical and Electronics Engineering, Sri Ramaswamy Memorial Institute of Science and Technology, Kattankulathur, India)

International Journal of Intelligent Unmanned Systems

ISSN: 2049-6427

Article publication date: 10 June 2019

Issue publication date: 13 June 2019



Humanoids have become the center of attraction for many researchers dealing with robotics investigations by their ability to replace human efforts in critical interventions. As a result, navigation and path planning has emerged as one of the most promising area of research for humanoid models. In this paper, a fuzzy logic controller hybridized with genetic algorithm (GA) has been proposed for path planning of a humanoid robot to avoid obstacles present in a cluttered environment and reach the target location successfully. The paper aims to discuss these issues.


Here, sensor outputs for nearest obstacle distances and bearing angle of the humanoid are first fed as inputs to the fuzzy logic controller, and first turning angle (TA) is obtained as an intermediate output. In the second step, the first TA derived from the fuzzy logic controller is again supplied to the GA controller along with other inputs and second TA is obtained as the final output. The developed hybrid controller has been tested in a V-REP simulation platform, and the simulation results are verified in an experimental setup.


By implementation of the proposed hybrid controller, the humanoid has reached its defined target position successfully by avoiding the obstacles present in the arena both in simulation and experimental platforms. The results obtained from simulation and experimental platforms are compared in terms of path length and time taken with each other, and close agreements have been observed with minimal percentage of errors.


Humanoids are considered more efficient than their wheeled robotic forms by their ability to mimic human behavior. The current research deals with the development of a novel hybrid controller considering fuzzy logic and GA for navigational analysis of a humanoid robot. The developed control scheme has been tested in both simulation and real-time environments and proper agreements have been found between the results obtained from them. The proposed approach can also be applied to other humanoid forms and the technique can serve as a pioneer art in humanoid navigation.



Rath, A.K., Parhi, D.R., Das, H.C., Kumar, P.B., Muni, M.K. and Salony, K. (2019), "Path optimization for navigation of a humanoid robot using hybridized fuzzy-genetic algorithm", International Journal of Intelligent Unmanned Systems, Vol. 7 No. 3, pp. 112-119.



Emerald Publishing Limited

Copyright © 2019, Emerald Publishing Limited

1. Introduction

With the development of science and technology, robots have become an integral part of daily life. Starting from medical assistance to industrial automation, various forms of robots are used extensively. However, humanoid robots have acquired a superior place owing to their ability to replace human efforts in repetitive tasks. To perform in a complex arena with maximum efficiency, humanoids require artificial intelligence (AI) to be integrated with their controller. As a result, several researchers have focused their attention toward design and implementation of smart algorithm for smooth navigation and motion planning of humanoid and other robotic forms. Some of the prominent research works in the above areas can be highlighted over here.

Patacchiola and Cangelosi (2017) have developed an approach to estimate the head pose of a robotic platform using a neural network technique. They have analyzed the working efficiency for wild data sets. Savage et al. (2013) have designed a genetic algorithm (GA)-based approach for navigational analysis of a mobile robot. Ming et al. (1996) have described about a fuzzy logic controller hybridized with GA in a biped robot. Using simulation tests, they have designed a control system for a semi-autonomous mobile robot to avoid obstacles present in an unknown environment. Kong et al. (2004) have used GA to study the gait analysis of a humanoid robot. They have designed the program using C++ in PC. Villela et al. (2017) have developed a stable and fast walking humanoid robot. They have performed the experiments using an NAO humanoid robot in a V-REP simulation platform. Son et al. (2017) have integrated a PID module with a feed forward neural network to control an SCARA robot’s joint positions. Tamayo et al. (2017) have analyzed and discussed the control methods of an AH1N2 humanoid robot. Rath, Das, Parhi and Kumar (2018), Rath, Parhi, Das and Kumar (2018) and Rath, Parhi, Das, Muni and Kumar (2018) have used AI techniques for optimizing the path of humanoid robot in an environment of disordered obstacles. They have performed both simulation and experimental analysis with good agreement between the results. Boukezzoula et al. (2018) have used fuzzy logic as a data integration and sensor fusion model for a humanoid robot. They have also used a camera input in their analysis. Pierezan et al. (2017) have discussed a modified self-adaptive differential evolution approach for humanoid robot with a series of experiments. Kumar et al. (2018a, b), Kumar, Rawat and Parhi (2018) and Kumar, Sethy and Parhi (2018) have studied and analyzed integration of smart algorithms in navigational control of humanoid robotic forms. Fakoor et al. (2016) have implemented a sensor integration-based navigational scheme for a humanoid robot. Abdolshah et al. (2017) have presented a humanoid walking model with the use of fuzzy logic and obtained satisfactory results. Lee et al. (2006) have presented gait analysis of a humanoid robot using fuzzy integrated GA. They have validated the proposed method through simulation analysis taking actual humanoid parameters. Nojima et al. (2003) have performed a series of experiments to generate human-like trajectories in humanoid robots and obtained suitable outcomes. Sahoo et al. (2018) and Rawat et al. (2018) have reviewed regarding use of various artificial intelligent algorithms for solving several engineering optimization problems. Samant et al. (2016) have proposed a fuzzy logic-based interaction model for a humanoid robot. Hassanzadeh and Sadigh (2009) have presented a fuzzy controller using GA for path optimization of a biped robot. MATLAB has been used for simulation analysis, and the experimental work has been performed using Khepera-II robot. Benbouabdallah and Qi-dan (2013) have proposed a target tracking approach of a mobile robot using soft computing technique. Di Nuovo et al. (2013) have developed a memory imagination-based control scheme for a humanoid robot.

The extensive literature survey indicates that most of the research works performed on humanoids are generally focused on control estimation. Navigational analysis for a humanoid robot in a complex arena with the presence of random obstacles is limited. Along with that, use of hybridized AI techniques in humanoid navigation is quite limited. Based on the limitations available with the existing literature citations, the current research is aimed at design and implementation of a novel hybrid fuzzy-GA for navigational analysis of a humanoid robot. The developed approach has been implemented on an NAO humanoid robot in both simulation and experimental environments and the results obtained from both the environments are compared against each other with proper agreement.

2. Analysis of hybridized fuzzy-genetic controller for a humanoid robot

With the need to solve variety of engineering optimization problems with better accuracy and speed, AI techniques have been developed over the last few decades. All the AI approaches used for the above-stated purpose have their own advantages and limitations. However, to combine advantageous properties from individual algorithms and provide more robustness to the problem-solving approach, hybridizations are attempted between the standalone algorithms.

Humanoid navigation is a challenging area of investigation that requires proper selection of navigational parameters and control. In the current work, a hybridized fuzzy-genetic approach has been designed and developed for navigational analysis of a humanoid robot in a cluttered environment. Fuzzy logic is a knowledge-based reasoning process that considers human-like intelligence to solve a problem. In the current work, NAO has been used as the humanoid platform. NAO is a medium-sized programmable humanoid designed and developed by Aldebaran Robotics Group, France. NAO is equipped with a large variety of sensory network that includes infrareds, sonars, force resistors, tactile sensors, inertial board, etc. Here, the two ultrasonic sensors present in the chest of NAO have been considered for the analysis.

Fuzzy logic works on the basis of input parameters, output parameters and sample rule base. The total sensory/detection range of the NAO’s ultrasonic sensors is separated to three parts and designated as front obstacle distance (FOD), left obstacle distance (LOD) and right obstacle distance (ROD). The inputs to the fuzzy controller are FOD, LOD, ROD and bearing angle (BA). BA indicates current heading angle of the humanoid. The output from the fuzzy controller is the turning angle (TA) that the humanoid requires to avoid an obstacle. Here, the threshold range of detection has been set as 35 cm. So, after the sensors detect and obstacle in the threshold range, the fuzzy controller is activated. The fuzzy logic inputs are the path length and angle between the humanoid robot and obstacles for calculation of the distance and angle by using data sensors:

(1) D F O D = S 0 + S 1 2 ,
(2) D R O D = S 1 + S 2 2 ,
(3) D L O D = S 2 + S 3 2 ,
where S0, S1, S2, S3, …, Si the values are given by the ith sensor; D the distance between robot and obstacles.

The fuzzy controller is fed with a vast sample rule base based on which it generates the output solution regarding the TA for the humanoid. Here, in the hybrid controller, TA (1) is generated as the intermediate solution that is again used as one of the inputs for the GA controller.

GA is a nature-inspired algorithm derived from natural genetics. It is based on the process of selection of better quality genes in each generation and discarding weak solutions. Here, a 12-bit chromosome-sized string has been selected as the solution of GA. Along with regular inputs such as FOD, LOD and ROD, the output of the fuzzy controller TA (1) are fed as inputs to the GA controller. Based on the input variables, GA first generates an initial solution for the TA. The initial solution is checked for optimal criteria with their fitness values. Here, the optimal criteria are set as 100 number of iterations of the improvement with each iteration being less than 2 percent of the previous solution. If optimal criteria are not met, an initial pool of solutions is generated as the parent chromosomes. The parent chromosomes then undergo for the process of crossover by exchanging their bits to generate better solutions as offspring. The offspring are then again checked for fitness valued and the better ones are added to the parent pool by discarding the weaker solutions. Crossover operation is also taken into account to add diversity in the population. The modified parent pool is again checked for optimality and the steps are repeated until the best solutions are achieved. The GA string calculates the fitness value as per the following convention:

(4) fitness i = α TOC i + β NOH i + γ fail i ,
where TOCi is the time of completion for ith solution (the time taken by the humanoid robot from start point to target point); NOHi is the number of hits of the robot with the obstacle for ith solution; faili the number of fails to reach the target for the humanoid robot for ith solution; α, β, γ are the constant coefficients.

The hybrid controller used in the current approach generates output of GA controller TA (2) as the final solution that guides the humanoid to avoid obstacles present in the arena. The total control process of the hybrid scheme can be summarized as follows. The humanoid starts navigating toward the goal from the source location with its initial heading angle set accordingly. Once the sensors detect an obstacle within the set threshold range, the hybrid controller gets activated. First, the fuzzy controller generates TA (1) as the intermediate output and then the GA controller generates the final output TA (2). The humanoid takes a turn according to TA (2) and then again heads toward the goal. The same procedure is repeated until it reaches the goal. Figure 1 represents the scheme of hybridization attempted in the current work.

3. Implementation of fuzzy-genetic hybrid controller in humanoid navigation

After formulating the control architecture of the fuzzy-genetic hybrid controller, it has been implemented for humanoid navigation in both simulation and experimental platforms. Here, V-REP has been used as the simulation platform. In a V-REP simulation platform, 250×240 units space has been designated as the navigation arena. Definite source and target locations have been marked for the humanoid. The simulation arena has been cluttered with five obstacles at random locations. The humanoid has been fed with the logic of the developed fuzzy-genetic hybrid controller, and navigational analysis has been performed. Figure 2 represents the simulation analysis results obtained from the analysis of humanoid navigation. It can be observed from Figure 2 that the humanoid has successfully reached the target location without colliding with the obstacles present in the arena.

To validate the simulation analysis results, the simulation arena conditions have been replicated in a real-time setup. An experimental platform with the same arena size as that of the simulation (250×200 units) has been developed under laboratory conditions. The location of source, target and obstacles has been kept exactly similar as that of simulation arena. In the experimental platform, the humanoid has been operated through a Wi-Fi controller. Plate 1 represents the results obtained from the analysis of humanoid navigation in experimental platform. The experimental analysis results also indicate that the proposed controller has generated satisfactory outcomes for collision-free smooth navigation of humanoid robot.

To compare the results obtained from simulation and experimental platforms, path length and time taken have been selected as the navigational parameters. They are recorded directly from the simulation window of V-REP software and are measured with the help of a measuring tape and stopwatch, respectively, from the experimental platform. Tables I and II represent the comparison of simulation and experimental results in terms of path length and time taken, respectively.

It can be observed from Tables I and II that the experimental results always generate higher values for the navigational parameters. The reason for the same can be explained by the presence of external factors such as data transmission loss, frictional loss, slippage effect, etc., in the experimental platform that are absent in case of the simulation arena as simulation arena operates in ideal conditions. However, the comparison of results also suggests a good agreement between the results and the maximum error limits observed are well within the acceptable range.

4. Conclusions

The current work has dealt with navigational analysis of a humanoid robot using a hybridized fuzzy-genetic controller in a complex arena. The developed controller has been designed in a two-step hybridization basis where the intermediate output from the fuzzy controller has been again fed as one of the inputs to the GA controller along with other regular inputs. The developed controller has been verified in a simulation platform, and the simulation results have been validated in an experimental setup prepared under laboratory conditions. The results obtained from the simulation and experimental platforms have been compared against each other and good agreement has been observed between them with minimal percentage of errors. The developed navigational technique can be used to generate smooth and hassle-free movement for a humanoid robot to operate with ease in planetary exploration, inspection of critical surroundings, hazardous environments, etc., which can be considered as the potential application areas of the current research.


Fuzzy-genetic hybrid architecture

Figure 1

Fuzzy-genetic hybrid architecture

Simulation analysis results

Figure 2

Simulation analysis results

Experimental analysis results

Plate 1

Experimental analysis results

Comparison of simulation and experimental path length

Sl. No. Simulation path length (m) Experimental path length (m) Deviation in path length (%) Average error (%)
1 3.15 3.33 5.4 5.58
2 3.05 3.3 7.57
3 3.02 3.2 5.62
4 3.17 3.34 5.08
5 3 3.13 4.15

Comparison of simulation and experimental time taken

Sl. No. Simulation time taken (s) Experimental time taken (s) Deviation in time taken (%) Average error (%)
1 55.6 57.2 2.79 4.28
2 52.5 55.2 4.89
3 49.2 51.6 4.65
4 53.9 56.8 5.1
5 50.1 52.2 4.02


Abdolshah, S., Abdolshah, M., Abdolshah, M. and Hashemi, S.V. (2017), “Walking control of humanoid robots on uneven ground using fuzzy algorithm”, Mathematical Concepts and Applications in Mechanical Engineering and Mechatronics, IGI Global, pp. 304-313, doi: 10.4018/978-1-5225-1639-2.ch014.

Benbouabdallah, K. and Qi-dan, Z. (2013), “Genetic fuzzy logic control technique for a mobile robot tracking a moving target”, International Journal of Computer Science Issues, Vol. 10 No. 1, pp. 607-613.

Boukezzoula, R., Coquin, D., Nguyen, T.L. and Perrin, S. (2018), “Multi-sensor information fusion: combination of fuzzy systems and evidence theory approaches in color recognition for the NAO humanoid robot”, Robotics and Autonomous Systems, Vol. 100, pp. 302-316.

Di Nuovo, A.G., Marocco, D., Di Nuovo, S. and Cangelosi, A. (2013), “Autonomous learning in humanoid robotics through mental imagery”, Neural Networks, Vol. 41, pp. 147-155.

Fakoor, M., Kosari, A. and Jafarzadeh, M. (2016), “Humanoid robot path planning with fuzzy Markov decision processes”, Journal of Applied Research and Technology, Vol. 14 No. 5, pp. 300-310.

Hassanzadeh, I. and Sadigh, S.M. (2009), “Path planning for a mobile robot using fuzzy logic controller tuned by GA”, 6th International Symposium on Mechatronics and its Applications, IEEE, March, pp. 1-5.

Kong, J.S., Lee, B.H. and Kim, J.G. (2004), “A study on the gait generation of a humanoid robot using genetic algorithm”, SICE 2004 Annual Conference, Vol. 1, August, pp. 187-191.

Kumar, P.B., Rawat, H. and Parhi, D.R. (2018), “Path planning of humanoids based on artificial potential field method in unknown environments”, Expert Systems, p. e12360, doi: org/10.1111/exsy.12360.

Kumar, P.B., Sahu, C. and Parhi, D.R. (2018a), “Navigation of humanoids by a hybridized regression-adaptive particle swarm optimization approach”, Archives of Control Sciences, Vol. 28 No. 3, pp. 349-378.

Kumar, P.B., Sahu, C. and Parhi, D.R. (2018b), “A hybridized regression-adaptive ant colony optimization approach for navigation of humanoids in a cluttered environment”, Applied Soft Computing, Vol. 68, pp. 565-585.

Kumar, P.B., Sethy, M. and Parhi, D.R. (2018), “An intelligent computer vision integrated regression based navigation approach for humanoids in a cluttered environment”, Multimedia Tools and Applications, Vol. 78 No. 9, pp. 11463-11486.

Lee, B.H., Kong, J.S. and Kim, J.G. (2006), “Optimal trajectory generation for a humanoid robot based on fuzzy and genetic algorithm”, IEEE Congress on Evolutionary Computation, IEEE, July, pp. 1968-1974.

Ming, L., Zailin, G. and Shuzi, Y. (1996), “Mobile robot fuzzy control optimization using genetic algorithm”, Artificial Intelligence in Engineering, Vol. 10 No. 4, pp. 293-298.

Nojima, Y., Kojima, F. and Kubota, N. (2003), “Trajectory generation for human-friendly behavior of partner robot using fuzzy evaluating interactive genetic algorithm”, Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation, Vol. 1, July, pp. 306-311.

Patacchiola, M. and Cangelosi, A. (2017), “Head pose estimation in the wild using convolutional neural networks and adaptive gradient methods”, Pattern Recognition, Vol. 71, pp. 132-143.

Pierezan, J., Freire, R.Z., Weihmann, L., Reynoso-Meza, G. and dos Santos Coelho, L. (2017), “Static force capability optimization of humanoids robots based on modified self-adaptive differential evolution”, Computers & Operations Research, Vol. 84, pp. 205-215.

Rath, A.K., Das, H.C., Parhi, D.R. and Kumar, P.B. (2018), “Application of artificial neural network for control and navigation of humanoid robot”, Journal of Mechanical Engineering and Sciences, Vol. 12 No. 2, pp. 3529-3538.

Rath, A.K., Parhi, D.R., Das, H.C. and Kumar, P.B. (2018), “Behaviour based navigational control of humanoid robot using genetic algorithm technique in cluttered environment”, Modelling, Measurement and Control A, Vol. 91 No. 1, pp. 32-36.

Rath, A.K., Parhi, D.R., Das, H.C., Muni, M.K. and Kumar, P.B. (2018), “Analysis and use of fuzzy intelligent technique for navigation of humanoid robot in obstacle prone zone”, Defence Technology, Vol. 14 No. 6, pp. 677-682.

Rawat, H., Parhi, D.R., Priyadarshi, B.K., Pandey, K.K. and Behera, A.K. (2018), “Analysis and investigation of Mamdani fuzzy for control and navigation of mobile robot and exploration of different AI techniques pertaining to robot navigation”, Emerging Trends in Engineering, Science and Manufacturing (ETESM-2018), IGIT, Sarang, March 28-29.

Sahoo, B., Parhi, D.R. and Priyadarshi, B.K. (2018), “Analysis of path planning of humanoid robots using neural network methods and study of possible use of other AI techniques”, Emerging Trends in Engineering, Science and Manufacturing (ETESM-2018), IGIT, Sarang, March 28-29.

Samant, R., Nair, S. and Kazi, F. (2016), “Development of autonomous humanoid robot control for competitive environment using fuzzy logic and heuristic search”, IFAC-Papers Online, Vol. 49 No. 1, pp. 373-378.

Savage, J., Muñoz, S., Matamoros, M. and Osorio, R. (2013), “Obstacle avoidance behaviors for mobile robots using genetic algorithms and recurrent neural networks”, IFAC Proceedings Volumes, Vol. 46 No. 24, pp. 141-146.

Son, N.N., Van Kien, C. and Anh, H.P.H. (2017), “A novel adaptive feed-forward-PID controller of a SCARA parallel robot using pneumatic artificial muscle actuator based on neural network and modified differential evolution algorithm”, Robotics and Autonomous Systems, Vol. 96, pp. 65-80.

Tamayo, A.J.M., Bustamante, P.V., Ramos, J.J.M. and Cobo, A.E. (2017), “Inverse models and robust parametric-step neuro-control of a humanoid robot”, Neurocomputing, Vol. 233, pp. 90-103.

Villela, L.F.C., Colombini, E.L., Técnico-IC-PFG, R. and de Graduação, P.F. (2017), “Humanoid robot walking optimization using genetic algorithms”, Projeto Final de Graduação, Julho.

Corresponding author

Asita Kumar Rath can be contacted at: