Satadal Ghosh

A study expands the SE-SCP algorithm and presents the generalized shape expansion (GSE) algorithm for a three-dimensional (3-D) environment. Contrary to the shape expansion performed by the SE-SCP, which is restricted to a spherical one with a small radius, the GSE algorithm helps in expansion over a generalized shape, which is the best representative of the free space in the overall workspace that helps in exploring the free space in a much more efficient way. That is why the word ‘generalized’ is used here. To this end, a sampling-based motion-planning algorithm has been presented, which explored a two-dimensional (2-D) workspace leveraging the novel GSE algorithm. It was found to explore the free space in a very efficient way, which was reflected in its computational advantage over several other existing seminal algorithms.

Localization of an unknown source emanating a radial monotonically varying scalar field in a planar environment by sensor-equipped unmanned aerial vehicles (UAVs) is considered in this paper. The source may be stationary, moving or maneuvering, but of constant signal strength. The only available information is the scalar measurements of the signal field taken by the on-board sensor of the UAV at each time instant. No range or bearing angle information is available to the UAV. Several methodologies have been explored for source seeking in the literature. Among them, a recently proposed algorithm leveraging an inter-loop switched-turn strategy, named 'Maxima Turn Strategy' (MTS), has been found to outperform several other strategies in terms of localization time-efficiency. However, the MTS strategy has a drawback in the sense that between two turn switching points the UAV traverses an entire circular arc, the radius of which corresponds to a pre-fixed nominal turn rate of the UAV. Thus, the MTS could be further improved upon by leveraging information about the gradient estimate updated at the end of the previous loop. Such a gradient-augmented MTS strategy that further reduces the intra-loop traversed path length with quite low memory storage requirement is presented in this paper. Effectiveness of the presented gradient-augmented MTS in terms of source localization time is validated by simulation studies with varied initial geometries for stationary, moving and maneuvering sources.

This paper presents an inter-loop switched-turn strategy to drive a mobile sensory aerial agent to the unknown source of a scalar signal field in planar and 3-D environments. The source may be stationary, moving or maneuvering, but assumed to be of constant signal strength. No range, bearing angle or gradient information is considered to be available. The only available information are the scalar measurements of the signal field taken by the on-board sensor of the UAV at each time-instant, and the signal strength is assumed to lessen with distance from the source. The effectiveness of the presented method in improving estimation of the direction to the source and in approaching closer to the source in each subsequent loop is shown, and verified by simulation results, both in the presence and absence of noise in signal measurement. The time-efficiency of the proposed method for varied initial conditions, and stationary and mobile sources, is shown through simulation results by comparison with two existing methods.

The problem of motion planning for an unmanned aerial vehicle (UAV) in an environment cluttered with stationary obstacles has found growing attention. To this end, several algorithms, both deterministic and sampling-based, have been reported in the literature. Recently, a sampling-based motion planning algorithm, named ’GSE’, has been presented for 2-D environment based on a novel sampling strategy-generalized shape expansion. Its effectiveness was found to be equivalent or even better in most aspects with respect to some existing seminal algorithms. Since for UAVs, 3-D environment is of more significance, this paper expands over the GSE algorithm and presents ’3D-GSE’ algorithm for generating shortest path and locally optimal trajectory from within the updated homotopy class using sequential convex programming in stationary obstacle-dense 3-D environments. The shape constructed by the 3D-GSE algorithm covers the available free space maximally, which is much more challenging than in case of 2-D environment. Thus, 3D-GSE exhibits its superiority in performance in terms of computational time efficiency when compared with well-established existing algorithms in extensive numerical simulation study. The trajectory costs obtained by the 3D-GSE algorithm is also found to be marginally better. With all these desired features, the 3D-GSE algorithm has a potential of real-time-implementation.

A major aspect of motion planning is the use of sampling-based algorithms. Sampling-based methods are primarily used to generate a feasible collision-free path for agents in an environment known a-priori. A recently proposed motion planning algorithm, termed as 'Generalized Shape Expansion' (GSE) algorithm, is a promising option in this class of algorithm. Extensive numerical studies have suggested that the GSE outperforms several seminal algorithms in literature in terms of computational time. However, so far no guarantee of probabilistic completeness of the GSE has been presented in literature. To this end, this paper elaborates a detailed mathematical analysis of GSE, providing upper bounds on the probability of failure of the GSE algorithm. A numerical example is presented to illustrate the proof. Simulation studies are presented to compare it with prominent algorithms in the literature, particularly in terms of number of iterations to reach a feasible path.

Precision soft landing guidance problem of a spacecraft on asteroid is addressed in this paper. Most of the methods in existing literature require linearized engagement dynamics or accurate time-to-go estimate, which is not easy to obtain in practice. To overcome these drawbacks, a heading error-dependent sliding-mode-control–based guidance approach was recently introduced in literature. However, it too suffered from disadvantages in terms of optimality and constant gravity setup-based formulation. To obviate these limitations, and, moreover, to accommodate the effect of asteroid’s angular rotation on the engagement dynamics, a sliding-mode-control– based instantaneously optimal (SMC-IO) guidance is presented in this paper, under a variable gravity setup, in which a heading-error-based sliding variable and a predefined range-dependent maximum velocity envelope profile are leveraged. A static optimization problem with square of instantaneous guidance command as the performance index and the sliding variable dynamics, velocity profile, equations of engagement, and thrust limit as constraints is solved to obtain the guidance command at every time instant. Simulation studies are presented to demonstrate the effectiveness of the presented SMC-IO guidance. An extensive comparative simulation study suggests superior performance of the SMC-IO guidance over widely referred existing landing guidance in terms of overall control effort and fuel usage.

An intelligent control algorithm based on adaptive sliding mode control and adaptive neural network for motion control of a quadcopter is presented in this paper. Full six degrees-of-freedom (6-DoF) model of a quadrotor with unknown parameters has been considered for this purpose. The ability of the presented control algorithm to work satisfactorily even in the absence of a-priori information of model parameters is its salient feature. Effectiveness of this algorithm in tracking nominal paths is illustrated by simulation results.

We present an online motion planning algorithm (3D-OGSSE) for generating smooth, collision-free trajectories over multiple planning iterations for a 3-D agent operating in an unknown, obstacle-cluttered, 3-D environment. In each planning iteration, 3D-OGSSE constructs an obstacle-free region termed 'generalized sensed shape' based on the locally-sensed environment information and the notion of generalized shape. A collision-free path is computed by sampling points in the generalized sensed shape and is used to generate a smooth, time-parametrized trajectory by minimizing snap. The generated trajectory at every planning iteration is constrained to lie within generalized sensed shape, which ensures the agent maneuvers in locally obstacle-free space. As the agent reaches the boundary of the generalized sensed shape in a planning iteration, a re-plan is triggered by a receding horizon planning mechanism that also enables the initialization of the next planning iteration. We also present a theoretical guarantee for probabilistic completeness of the developed algorithm over the entire environment and for completely collision-free trajectory generation. We evaluate the proposed method in simulation on complex 3-D environments with varied obstacle-densities. Further, we also evaluate it in scenarios with sensor noise and constraints on the on-board sensor's field-of-view (FOV). We observe that each planning iteration computation takes $\sim 14$ milliseconds on a single thread of an Intel Core i5-8500 3.0 GHz CPU, which is significantly faster than several existing algorithms. In addition, we also observe 3D-OGSSE to be less conservative in complex scenarios such as narrow passages.

Optimal tracking of continuous-time non-linear systems has been extensively studied in literature. However, in several applications, absence of knowledge about system dynamics poses a severe challenge in solving the optimal tracking problem. This has found growing attention among researchers recently, and integral reinforcement learning based method augmented with actor neural network (NN) have been deployed to this end. However, very few studies have been directed to model-free H∞ optimal tracking control that helps in attenuating the effect of disturbances on the system performance without any prior knowledge about system dynamics. To this end, a recursive least square-based parameter update was recently proposed. However, gradient descent-based parameter update scheme is more sensitive to real-time variation in plant dynamics. Experience replay (ER) technique has been shown to improve the convergence of NN weights by utilising past observations iteratively. Motivated by these, this study presents a novel parameter update law based on variable gain gradient descent and ER technique for tuning the weights of critic, actor and disturbance NNs. The presented update law leads to improved model-free tracking performance under ℒ2-bounded disturbance. Simulation results are presented to validate the presented update law.

Due to the rapid increase of unmanned aerial vehicle (UAV) usage, the demand for efficient autonomous interdiction techniques to safeguard protected areas has become increasingly essential. This paper presents novel guidance strategies based on Proportional Navigation (PN) to interdict a stationary target using single and multiple unmanned aerial vehicles (UAVs). While the previous literature has primarily addressed controlling the terminal angle and achieving a desired final time separately for single-pursuer and multi-pursuer setups, designing guidance strategies to achieve both simultaneously poses a significant challenge. Although few existing literature endeavor to satisfy both constraints, they lack in guaranteeing an all-aspect approach. To this end, this paper's main contribution is enabling pursuers to achieve any terminal configuration starting from any initial orientation while satisfying the final time constraint by employing PN-based multi-phase guidance strategies in single and multiple pursuer setups. While the 'Preparation phase' at the beginning and the 'Final PPN phase' at the end help ensure the desired terminal orientation, the intermediate Roaming phase helps achieve the desired final time. Also, the guarantee on phase transitions and performance of the overall guidance schemes and conditions on achievable final time for the success of the developed guidance schemes are analyzed. Finally, using numerical simulations, the developed guidance algorithms are validated for single and multiple pursuer(s) environments considering realistic constraints.