Hybrid formation control of multiple mobile robots with obstacle avoidance

A survey of formation control and motion planning of multiple unmanned vehicles

“…Since it is hard to mathematically analyse the formation stability by using the behaviour-based method, a hybrid control scheme that includes both the leader-follower and the behaviour-based methods was proposed by Yang et al 61 The formation was generated and maintained by the leaderfollower while the behaviour-based scheme specifically focused on the motion planning of individual vehicles. A supervision mechanism has been built between the leader and followers such that the formation integrity can be ensured when the number of controlled vehicles changes.…”

Distributed formation control of networked mobile robots in environments with obstacles

“…Coupled with the small output from the VC, the robot is unable to move to its FP. To overcome this, a wall following technique based on [19] has been implemented. When an obstacle is detected within R wf and sits beyond a threshold angle, Θ side (measured from the centre of the forward arc), the robot is forced to travel in the forward direction at V lock , which a sufficiently small value which allows it to navigate around tight corners.…”

A deep learning optimized LQR method for enhanced formation control with embedded systems

“…These formation methodologies necessitate a stable communication infrastructure and exhibit a pronounced dependency on the central robot. In contrast, within the purview of the latter methodologies, the formation system endows each individual robot with the capability to autonomously perceive the environment and the formation structure, as well as to execute control actions independently (Yang et al 2007). In a further step, the formation control can be classified into several types, such as the leader-follower methods (Defoort et al 2008, Gao and, the virtual structure methods (Lewis andTan 1997, Sadowska et al 2011), the artificial field methods (Khatib 1986, Zhu et al 2010), the behavior-based methods (Cristoforis et al 2013, Liang et al 2013, the graph theory methods (Yang et al 2014), and so on.…”

“…Among these approaches are control laws grounded in Lyapunov theory, specifically tailored to govern the actions and behaviors of follower vehicles (Zhang et al 2014), A displacement-based formation control algorithm was designed to achieve the disturbance rejection and was verified in simulation (Aldarmini et al 2022). However, for more practical applications, direct tracking control methods were adopted such as the conventional PID control (Yang et al 2007), the artificial potential field method that generated repulsive forces from obstacles and attractive force from goals (Jia et al 2006, Toyota andNamerikawa 2017), the geometry-based Stanley which considered the real-time velocity (Thrun et al 2006), the pure-pursuit control method in which the simplified kinematic model was used to compute steering angle for the robot (Elbanhawi et al 2018), the neural adaptive PID control which used a neural network to dynamically generate the parameters of the PID controller (Shojaei 2017), and the Dynamic Window Approach (DWA) method which estimated the predictive trajectories while in formation (Ling et al 2019). The vision-based formation control framework was also studied by using only one omnidirectional camera mounted on the top of each robot (Das et al 2002).…”