A Review on Path Planning and Obstacle Avoidance Algorithms for Autonomous Mobile Robots

Table 2

Summary of a few works performed using the heuristic approach.

Author	Approach	Objective	Contribution
Moor and Vachtsevanos [44]	Fuzzy logic	The authors started using the idea of FL for robotic path planning with obstacle avoidance.	This study aims to offer a method for determining the shortest path for a quadrotor to navigate in an unfamiliar environment.
Wang and Liu [48]	Fuzzy logic	The authors introduced the method of FL path planning in an unknown environment.	This paper proposes a novel technique, the minimal risk approach, for tackling the local path planning problem in goal-oriented robot navigation in unknown environments to avoid the local minimum.
Xiang et al. [52]	Fuzzy logic	The navigation in 3-dimension space was presented for underwater robots.	This research aims to see how well fuzzy-logic-based guiding works in one of the most important areas of robotics: Marine robotic vehicles. Recent improvements in fuzzy control include self-tuning, direct/indirect adaptive fuzzy control, and neuro-fuzzy control.
Zhang, et al. [63]	Neural network	We utilized mobile robot path planning problems for aerial robots.	The recommended controller directly generates angular velocity commands using the outer position loop. Simulations and actual flight tests are used to validate the feasibility and efficacy of the suggested control approach.
Zhu et al. [65]	Neural network	The Glasius bio-inspired neural network (GBNN) method is presented to enhance the path planning of autonomous underwater vehicles.	An underwater grid map is constructed by discretizing the two-dimensional underwater environment. Then, a dynamic neural network is built on top of the grid map. As a consequence, AUV can cover the workspace and avoid deadlocks.
Sun, He and Hong [66]	Neural network	Adaptive neural networks (NNs) are used in the control design of a flexible robotic manipulator.	The system is modeled using the lumped spring-mass method. Full-state feedback control and output feedback control are also offered.
Kumar et al. [70]	Particle swarm optimization	Adaptive neural networks (NNs) are used in the control design of a flexible robotic manipulator.	The system is modeled using the lumped spring-mass method. Full-state feedback control and output feedback control are also offered. Experiments are conducted to validate further the viability of the suggested NN controllers on the Quanser platform.
Rendom and Martins [73]	Particle swarm optimization	To propose a particle swarm optimization application for adjusting quadrotor attitude and route following control.	Using Euler–Lagrange equations, path planning will be carried out to reduce the snap cost function and ensure a smooth trajectory. Several simulations will be used to assess the dependability of this technique.
Patle et al. [92]	Genetic algorithm	To demonstrate SLI via accurate and practical data for the assertion.	The suggested controller’s result is optimal in terms of path and time when compared to existing intelligent navigational controllers.
Kumar et al. [95]	Genetic algorithm	Navigation in the presence of moving obstacles in an uncertain environment.	It was determined that the suggested navigational controllers are efficient in path planning and obstacle avoidance and may be used for humanoid navigation in complicated settings.
Mohanty and Parhi [106]	Cuckoo search algorithm	The authors proposed a stand-alone CSA to navigate the wheeled MR in a static, partially unknown environment.	The results of simulations demonstrate that this technique may provide a safe and effective path design.
Contreras-Cruz et al. [111]	Artificial bee colony	To define the optimal path in real-time.	The suggested technique is compared to a traditional probabilistic roadmap method (PRM) in planning performance on a set of benchmark issues, outperforming it.
Ding et al. [119]	Artificial bee colony	To obtain an optimal path in 3D world and unmanned helicopter for undergoing the challenge mission, etc.	The identification findings using input-output data from actual flying tests demonstrated the superiority of CABC over the ABC and the genetic algorithm (GA).