DESC:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:

METHOD AND SYSTEM FOR DECENTRALIZED SAFE NAVIGATION IN MULTI-ROBOTIC AGENT SYSTEMS

Applicant:
Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th Floor,
Nariman Point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the invention and the manner in which it is to be performed.
CROSS-REFERENCE TO RELATED APPLICATIONS AND PRIORITY
[001] The present application claims priority from Indian provisional patent application no. 202421034679, filed on May 01, 2024. The entire contents of the aforementioned application are incorporated herein by reference.

TECHNICAL FIELD
[002] The disclosure herein generally relates to multiple robotic agents’ navigation, and, more particularly, to a method and system for decentralized safe navigation in multi-robotic agent systems.

BACKGROUND
[003] Technological progress in distributed sensing and computational capabilities has facilitated the integration of robots into diverse environments, including but not limited to autonomous transportation, multi-robot task allocation in warehouse environments, dynamic coverage, inspection, and cleaning services. Ensuring the safety of robots in safety-critical applications, particularly when navigating to a goal position or a target point while adhering to actuation and safety constraints, presents a significant challenge. Furthermore, the real-world operation of many such systems is susceptible to disturbances, leading to a growing focus on formal verification of autonomous system safety in recent years.
[004] In the realm of safe control navigation, two fundamental objectives take precedence: (a) achieving reachability to a designated goal set and (b) ensuring the forward invariance of a predefined safe set. Traditionally, addressing the safe control objectives concurrently involves employing a combination of Control Lyapunov Functions (CLFs) and Control Barrier Functions (CBFs), often through the application of quadratic programming (QP)-based control synthesis. However, manually synthesizing CBFs for intricate dynamic systems proves to be an exceedingly challenging task. Nonetheless, it is crucial to note that existing studies in the domain of learning-based approaches for CBF-based designs have predominantly focused on single-robotic systems. Moreover, other approaches that rely on learning CBFs using data-driven approaches do not come with formal safety guarantees.
[005] Thus, the extension of learning-based approaches to enable safe control in multi-robotic agent systems, while maintaining both provable dependability and scalability, remains an open and unexplored area within the field. In, managing multi- robotic agent systems introduce distinct challenges, primarily centered around achieving effective coordination among the multiple robotic agents. A centralized approach offers the promise of flawless coordination but quickly becomes computationally infeasible when applied to a larger number of multiple robotic agents. Conversely, a decentralized approach, while scalable, often encounters coordination and performance challenges. Consequently, there has been a growing inclination toward embracing data-driven methodologies for safe control. Nevertheless, the overarching challenge of ensuring the broad applicability and generalizability of these learning-based approaches persists.

SUMMARY
[006] Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems. For example, in one aspect, a method for decentralized safe navigation in multi-robotic agent systems is provided. The method comprises: collecting, via a robotic agent, real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent; determining, via the robotic agent, a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent by processing the real-time on-board sensor data.
[007] Further, the method comprises: iteratively performing, via the robotic agent, a sequence of steps to determine a navigation path towards a goal position allocated to the robotic agent, wherein the sequence of steps, executed in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent. The sequence of steps comprising: identifying a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent; modelling an optimization problem using a Control Lyapunov Fields (CLF-QP) technique based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position; modelling a set of non-convex elements in the optimization problem and a set of collision avoidance constraint using a Sequential Convexification Programming (SCP), wherein SCP allows convexification of the optimization problem; adding a random zero mean Gaussian noise to the current position of the robotic agent, the current position of each of the plurality of robotic agents; and generating, an optimal control action at the time instant for of the robotic agent in a decentralized manner by solving the optimization problem; executing, the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant; and modifying, the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius, wherein the robotic agent navigates towards the goal position with each iteration.
[008] Further the method establishes a safe corridor traversal for the modified navigation path through a confined corridor using a safe corridor traversal technique, for achieving decentralized safe navigation of the robotic agent within a dynamic environment.
[009] Further, the method for the safe corridor traversal comprises iteratively performing a sequence of steps until the robotic agent attains the goal position, the sequence of steps comprising: dividing the goal position of each of the plurality of robotic agents into a plurality of sub-goal regions; defining, one or more elliptical regions for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions; setting, the goal position of the robotic agent to a centroid of the intersection of first and second elliptical regions; and achieving, the goal position by the robotic agent using the modified navigation path. The method further includes shifting the goal position of the robotic agent to the centroid of the second and third elliptical regions, till each of the plurality of robotic agents reaches the goal positions allocated.
[010] In another aspect, a system (robotic agent) for decentralized safe navigation in multi-robotic agent systems is provided. The system comprises a memory storing instructions; one or more Input/Output (I/O) interfaces; and one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to collect real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent; determine a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent by processing the real-time on-board sensor data.
[011] The system (robotic agent) is configured to iteratively perform a sequence of steps to determine a navigation path towards a goal position allocated to the robotic agent, wherein the sequence of steps, executed in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent. The sequence of steps comprising: identifying a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent; modelling an optimization problem using a Control Lyapunov Fields (CLF-QP) technique based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position; modelling a set of non-convex elements in the optimization problem and a set of collision avoidance constraint using a Sequential Convexification Programming (SCP), wherein SCP allows convexification of the optimization problem; adding a random zero mean Gaussian noise to the current position of the robotic agent, the current position of each of the plurality of robotic agents; and generating, an optimal control action at the time instant for of the robotic agent in a decentralized manner by solving the optimization problem; executing, the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant; and modifying, the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius, wherein the robotic agent navigates towards the goal position with each iteration.
[012] Further, the system is configured to establish a safe corridor traversal for the modified navigation path through a confined corridor using a safe corridor traversal technique, for achieving decentralized safe navigation of the robotic agent within a dynamic environment. The system is further configured to iteratively perform a sequence of steps until the robotic agent attains the goal position, the sequence of steps comprising: dividing the goal position of each of the plurality of robotic agents into a plurality of sub-goal regions; defining, one or more elliptical regions for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions; setting, the goal position of the robotic agent to a centroid of the intersection of first and second elliptical regions; and achieving, the goal position by the robotic agent using the modified navigation path. The method further includes shifting the goal position of the robotic agent to the centroid of the second and third elliptical regions, till each of the plurality of robotic agents reaches the goal positions allocated.
[013] In yet another aspect, there are provided one or more non-transitory machine-readable information storage mediums comprising one or more instructions, which when executed by one or more hardware processors causes a method for decentralized safe navigation in multi-robotic agent systems.
[014] The method includes collecting real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent; determining a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent by processing the real-time on-board sensor data; and iteratively performing a sequence of steps to determine a navigation path towards a goal position allocated to the robotic agent, wherein the sequence of steps, executed in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent. The sequence of steps comprising: identifying a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent; modelling an optimization problem using a Control Lyapunov Fields (CLF-QP) technique based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position; modelling a set of non-convex elements in the optimization problem and a set of collision avoidance constraint using a Sequential Convexification Programming (SCP), wherein SCP allows convexification of the optimization problem; adding a random zero mean Gaussian noise to the current position of the robotic agent, the current position of each of the plurality of robotic agents; and generating, an optimal control action at the time instant for of the robotic agent in a decentralized manner by solving the optimization problem; executing, the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant; and modifying, the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius, wherein the robotic agent navigates towards the goal position with each iteration.
[015] In the non-transitory computer readable medium, a safe corridor traversal for the modified navigation path through a confined corridor is established using a safe corridor traversal technique, for achieving decentralized safe navigation within a dynamic environment.
[016] Further, the method comprises iteratively performing a sequence of steps until the robotic agent attains the goal position, the sequence of steps comprising: dividing the goal position of each of the plurality of robotic agents into a plurality of sub-goal regions; defining, one or more elliptical regions for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions; setting, the goal position of the robotic agent to a centroid of the intersection of first and second elliptical regions; and achieving, the goal position by the robotic agent using the modified navigation path.
[017] In another embodiment of the non-transitory computer readable medium, the goal position of the robotic agent is shifted to the centroid of the second and third elliptical regions, till each of the plurality of robotic agents reaches the goal positions allocated.
[018] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS
[019] The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles:
[020] FIG. 1 illustrates an exemplary block diagram of a system for decentralized safe navigation in multi-robotic agent systems, according to some embodiments of the present disclosure.
[021] FIG. 2A and FIG. 2B illustrate a method for decentralized safe navigation in multi-robotic agent systems, according to some embodiments of the present disclosure.
[022] FIG. 3 illustrates an intersection centroid approximation for a safe corridor traversal of a robotic agent in accordance with some embodiments of the present disclosure.
[023] FIG. 4 illustrates a safe corridor traversal of a robotic agent in accordance with some embodiments of the present disclosure.
[024] FIG. 5A through FIG.5F illustrates navigation of a multi-robotic agent system in a dynamic environment in accordance with some embodiments of the present disclosure.

DETAILED DESCRIPTION OF EMBODIMENTS
[025] Exemplary embodiments are described with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the scope of the disclosed embodiments.
[026] Conventionally, managing multi-robotic agent systems introduces distinct challenges, primarily centered around achieving effective coordination among diverse agents. A centralized approach offers flawless coordination among the agents in a multi-robotic agent system but immediately becomes computationally infeasible when applied to a system having large number of agents. On the other hand, a decentralized approach, while scalable, often encounters coordination and performance challenges. Thus, there has been a growing inclination towards adopting data-driven methodologies for safe control. Nevertheless, the overarching challenge of ensuring the broad applicability and generalizability of these learning-based approaches persists.
[027] Few prior works have designed Reinforcement Learning (RL) based algorithm for task allocation and decentralized navigation among multiple mobile robots. Such algorithms were not so useful for safety-critical applications, especially, when guiding mobile robots through a confined space with a designated goal set. Few prior works have strived to build a swarm-based approach for acquiring dynamic coverage control in multi robotic systems but the real-world operation of many such systems is prone to disturbances, making safe navigation a challenge.
[028] Thus, while considering safe and controlled navigation, achieving reachability to a designated goal set and ensuring the forward invariance of a predefined safe set are two essential objects in a multi robotic system. Conventionally, these objects are delivered using a combination of Control Lyapunov Functions (CLFs) and Control Barrier Functions (CBFs), along with the application of quadratic programming (QP)-based control synthesis. However, manually synthesizing CBFs for complex and dynamic multi-robotic agent systems requires immense amount of time and is extremely challenging in nature. Hence, driving the exploration of learning-based approaches for CBF-based design. Nonetheless, it is crucial to note that existing studies in the domain of CBF-based designs have predominantly focused on single-agent systems only. Enabling safe control in multi-robotic agent systems, while maintaining both provable dependability and scalability, remains as an unsolved problem.
[029] To overcome the above-mentioned drawbacks of the existing methods, the embodiment of the present disclosure provides a method for decentralized safe navigation in multi-robotic agent systems. In the context of the present disclosure, the expressions ‘mobile robotic agents and ‘robotic agents’ may be used interchangeably. The decentralized safe navigation is achieved by utilizing a combination of Control Lyapunov Function based quadratic programs (CLF-QP), sequential convex programming (SCP), and explicit safety constraints. The explicit safety constraints include constraints such as a safe radius directly fed into the CLF-QP in order to maintain a minimum safety distance by the robotic agents from other static or dynamic obstacles. The proposed method of decentralized safe navigation enables the robotic agents to autonomously navigate towards the designated goal sets while actively avoiding collisions with other robotic agents and also with static obstacles.
[030] Forward invariance to safety sets is guaranteed by creating a sequence of goal sets designed to navigate through exceedingly narrow safe regions. Furthermore, in contrast to conventional decentralized approaches, the proposed method operates without the need for access to the dynamics of other robotic agents.
[031] Since the decentralized safe navigation method relies solely on measurements of other robotic agents’ positions and velocities within a limited sensing field, adaptation of the method is both feasible and practical. Also, the proposed method is applicable to a wide range of nonlinear control-affine systems, including unmanned aerial vehicles (UAVs), differential drive robots, and more. Moreover, the proposed method ensures optimal scalability, even when dealing with a considerable number of robotic agents. Each robotic agent has a designated goal set while maintaining a safe distance from other moving robotic agents or dynamic obstacles and static obstacles. The robotic agents sometimes, in order to reach the designated goal position have to take a confined, narrow path of navigation. The proposed method enables a robotic agent to generate an optimal control action, decentralized in nature, so that the robotic agent or agents attains the goal position via a collision free navigation path.
[032] In the disclosed method a CLF-QP framework is used for each of the robotic agent in a control-affine system capable of reaching the goal set in a fixed-time independent of initialization. One notable advantage of the proposed method is the ability to seamlessly connect multiple goal sets, allowing the multi-robotic agent system to traverse through exceedingly narrow safe regions while simultaneously maintaining precise heading control. Further, the proposed method also achieves safe control even under limited sensing. In contrast to the idealized scenario of having complete and precise knowledge of both static and dynamic obstacles’ states, the proposed method operates within a more pragmatically relevant context, where obstacles can only be detected within a limited radius around the robotic agent. Moreover, the proposed method employs Sequential Convex Programming (SCP) for convexification of non-convex constraints arising from nonlinear nature of dynamical equations governing robotic agent navigation. The SCP approach helps in converting such non-linearities amenable to the proposed CLF-QP framework.
[033] Referring now to the drawings, and more particularly to FIG. 1 through FIG. 5F, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments, and these embodiments are described in the context of the following exemplary system and/or method.
[034] FIG. 1 illustrates an exemplary block diagram of decentralized safe navigation in multi-robotic agent systems, according to some embodiments of the present disclosure. In an embodiment, the system 100, also referred to as the robotic agent (100) includes one or more processors 104, communication interface device(s) 106 or Input/Output (I/O) interface(s) 106 or user interface 106, and one or more data storage devices or memory 102 operatively coupled to the one or more processors 104. The memory 102 comprises a database 108. The one or more processors 104 that are hardware processors can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor(s) is configured to fetch and execute computer-readable instructions stored in the memory. In an embodiment, the system 100 can be implemented in a variety of computing systems, such as laptop computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud, and the like.
[035] The I/O interface device(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite. In an embodiment, the I/O interface device(s) 106 receives a material characteristic data and a user query as input and gives recommendation of matching materials as output.
[036] The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random-access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.
[037] In an embodiment, the memory 102 includes a plurality of modules 110 for realizing decentralized safe navigation in multi-robotic agent systems. The plurality of modules 110 include programs or coded instructions that supplement applications or functions performed by the system 100 for executing one or more steps involved in the method of decentralized safe navigation in multi-robotic agent system. The plurality of modules 110, amongst other things, can include routines, programs, objects, components, and data structures, which perform particular tasks or implement particular abstract data types. The plurality of modules 110 may also be used as, signal processor(s), node machine(s), logic circuitries, and/or any other device or component that manipulates signals based on operational instructions. Further, the plurality of modules 110 can be used by hardware, by computer-readable instructions executed by the one or more hardware processors 104, or by a combination thereof. The plurality of modules 110 can include various sub-modules (not shown).
[038] Functions of the components of system 100 are explained in conjunction with flow diagram depicted in FIG. 2A for decentralized safe navigation in multi-robotic agent systems.
[039] In an embodiment, the system 100 comprises one or more data storage devices or the memory 102 operatively coupled to the processor(s) 104 and is configured to store instructions for execution of steps of the method (200) depicted in FIG. 2A by the processor(s) or one or more hardware processors 104. The steps of the method of the present disclosure will now be explained with reference to the components or blocks of the system 100 as depicted in FIG. 1 and the steps of flow diagram as depicted in FIG. 2A. Although process steps, method steps, techniques or the like may be described in a sequential order, such processes, methods, and techniques may be configured to work in alternate orders. In other words, any sequence or order of steps that may be described does not necessarily indicate a requirement that the steps to be performed in that order. The steps of processes described herein may be performed in any order practical. Further, some steps may be performed simultaneously.
[040] FIG. 2A is a flow diagram illustrating method (200) of a decentralized safe navigation for one or more robotic agents within a dynamic environment, according to some embodiments of the present disclosure. In the method 200, one or more hardware processors 104 process a real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent. At step 202 of the method 200, the one or more hardware processors 104, are configured by the instructions to: collect the real time onboard sensor data from real-time onboard sensors such as LIDAR, sonars etc. At step 204 of the method 200, the one or more hardware processors 104, are configured by the instructions to determine a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent from the processed real-time on-board sensor data.
[041] At step 206 of the method 200, the one or more hardware processors 104, are configured by the instructions to: iteratively perform a sequence of steps to determine a navigation path towards a goal position allocated to the robotic agent. The sequence of steps is explained in conjunction with flow diagram FIG. 2B from 206a to 206g, executed is in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent under consideration.
[042] At step 206a of the method 200, the one or more hardware processors 104 are configured by the instruction to identify a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent. Each of the plurality of robotic agents are assigned with a designated goal position. Further at the step 206a, each of the plurality of robotic agent navigates towards the designated goal set while maintaining a safe distance, called the safe radius, from the static and dynamic obstacles. The dynamic obstacles include but not limited to the plurality of robotic agents near to the robotic agent under consideration.
[043] At step 206b, of the method 200, an optimization problem using a Control Lyapunov Fields (CLF-QP) technique is modelled. Also, the optimization problem modelling is based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position. Further, at step 206c of the method 200, a set of non-convex elements in the optimization problem and a set of collision avoidance constraint is modelled using a Sequential Convexification Programming (SCP). In the aspect of the present disclosure the SCP allows convexification of the optimization problem.
[044] At step 206d, of the method 200, a random zero mean Gaussian noise is added to the current position of the robotic agent as well as to the current position of each of the plurality of robotic agents. The random zero mean Gaussian noise added to the current position addresses symmetric deadlocks in the optimization problem. The symmetric deadlocks arises when the robotic agent and any one of the pluralities of the robotic agents have same navigation path towards the goal position. The random zero mean Gaussian noise helps breaking the symmetry in the navigation path, by varying the control action output, while maintaining the safe radius between the plurality of robotic agents.
[045] Further, at step 206e, of the method 200, the optimization problem is solved in order to generate an optimal control action at the time instant for the robotic agent. In the aspect of the present disclosure, the optimal control action is one of (i) a symmetric and (ii) non-symmetric, wherein the symmetric optimal control action generated allows the robotic agent to obtain a symmetrical path, while the non-symmetric optimal control action generated allows the robotic agent to break the symmetrical path when the static or dynamic obstacle is in the symmetrical path. Thus, the optimal control action generated is decentralized in nature for each of the plurality of the robotic agents in a multi-robotic agent system.
[046] At step 206f, of the method 200, the one or more hardware processors 104, are configured by the instructions to execute the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant Further, at step 206g, of the method 200, the one or more hardware processors 104 are configured by the instructions to modify the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius.. Thus, the robotic agent navigates towards the goal position with each iteration of generating the optimal control action.
[047] At step 208, of the method 200, a safe corridor traversal for the modified navigation path through a confined corridor is established, as shown in FIG. 4 using a safe corridor traversal technique. The safe corridor traversal is realized using an intersection centroid approximation as depicted in FIG. 3 based on the modified navigation path of the robotic agent. The intersection centroid approximation is a simple approach to find the centroid of the intersection of two overlapping elliptical regions (safe sets).
[048] The safe corridor traversal technique includes the following steps. At first the goal position of each of the plurality of robotic agents are divided into sub-goal regions. Using the intersection centroid approximation, the centroid of the overlapping sub-goal regions is approximated as the intersection of the diagonals of a quadrilateral formed by the tangents drawn to both ellipses at the points of intersection. Thus, one or more elliptical regions are defined for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions.
[049] The goal position of the robotic agent is set to centroid of the intersection of first and second elliptical regions and the robotic agent traverse towards the goal position using the modified navigation path. Thus, the safe corridor is approximated as a sequence of intersecting ellipses, each intersecting at exactly two distinct points. The navigation path of the robotic agent remains confined within this safe corridor throughout the traversal. After attaining the goal position, the robotic agent enters the second elliptical region. The goal position of the robotic agent is then shifted to the centroid of the second and third elliptical regions. Finally, at step 208 of the method 200, the steps are repeated for next set of elliptical regions till each of the plurality of robotic agents reaches the goal positions allocated. Schematic representing the safe corridor traversal of the robotic agent by generating successive overlapping ellipses contained completely inside the corridor is depicted in the FIG. 4. The robotic agent thus realizes the safe corridor traversal towards for achieving decentralized safe navigation within a dynamic environment.
[050] FIG 5A to 5F illustrates the application of the proposed decentralized safe navigation framework for a multi-robotic agent system in a densely populated and dynamic environment featuring 15 robotic agents. The FIGS 5A, 5B, 5C, 5D, 5E and 5F depict sequential evolution of robotic agents’ navigation starting at initial positions and heading towards the associated goal positions while avoiding collisions with other mobile robotic agents appearing within the sensing field of each of the robotic agents. The proposed approach effectively guides each robot towards its respective goal while ensuring the safe avoidance of others by the specified avoidance threshold, highlighting the scalability and ease of implementation.
[051] The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments. The scope of the subject matter embodiments is defined by the claims and may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope of the claims if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language of the claims.
[052] The present disclosure describes a method for decentralized safe navigation in multi-robotic agent systems using a combination of Control Lyapunov Functions (CLFs) and Sequential Convex Programming (SCP) along with the application of quadratic programming (QP)-based control optimization. The embodiment of present disclosure herein addresses unresolved problem of safe and collision free navigation among multiple mobile robotic agents using a decentralized control within a dynamic environment unlike the prior arts which uses a combination of CLFs or Control Barrier Functions (CBFs) for a single robotic agent navigation. The embodiment of the present disclosure simultaneously prioritizes achieving controllability i.e., convergence to a set of goal position and safety, ensuring collision avoidance with both static and dynamic obstacles among multiple mobile robotic agents by incorporating SCP.
[053] The embodiment of the present disclosure results in a computationally efficient Control Lyapunov Functions- Sequential Convex programming (CLF-SCP) based Quadratic Programming (QP) solution, which holds the potential to scale seamlessly to accommodate an arbitrary number of agents. Moreover, the embodiments herein further achieve the scalability without the need for explicit knowledge of the dynamics of each of the robotic agents in the environment, eliminating the complexity associated with designing intricate barrier functions for safe control or relying on data-driven methods that may struggle with generalization. The embodiment thus provides a method for decentralized and collision free navigation for multiple robotic agents for safety critical applications.
[054] It is to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g., any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g., hardware means like e.g., an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g., an ASIC and an FPGA, or at least one microprocessor and at least one memory with software processing components located therein. Thus, the means can include both hardware means and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments may be implemented on different hardware devices, e.g., using a plurality of CPUs.
[055] The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various components described herein may be implemented in other components or combinations of other components. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
[056] The illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope of the disclosed embodiments. Also, the words “comprising,” “having,” “containing,” and “including,” and other similar forms are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise.
[057] Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present disclosure. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, nonvolatile memory, hard drives, CD ROMs, DVDs, flash drives, disks, and any other known physical storage media.
[058] It is intended that the disclosure and examples be considered as exemplary only, with a true scope of disclosed embodiments being indicated by the following claims.

,CLAIMS:
1. A processor implemented method (200) for achieving a decentralized safe navigation for one or more robotic agents within a dynamic environment, the method comprising:
collecting (202), via a robotic agent controlled by one or more hardware processors, real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent;
determining (204), via the robotic agent controlled by one or more hardware processors, a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent by processing the real-time on-board sensor data;
iteratively performing, via the robotic agent controlled by one or more hardware processors, a sequence of steps to determine (206) a navigation path towards a goal position allocated to the robotic agent, wherein the sequence of steps, executed in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent, the sequence of steps comprising:
identifying (206a) a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent;
modelling (206b) an optimization problem using a Control Lyapunov Fields (CLF-QP) technique based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position;
modelling (206c) a set of non-convex elements in the optimization problem and a set of collision avoidance constraint using a Sequential Convexification Programming (SCP), wherein SCP allows convexification of the optimization problem;
adding (206d) a random zero mean Gaussian noise to the current position of the robotic agent, the current position of each of the plurality of robotic agents; and
generating (206e), an optimal control action at the time instant for of the robotic agent in a decentralized manner by solving the optimization problem;
executing (206f), the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant; and
modifying (206g), the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius, wherein the robotic agent navigates towards the goal position with each iteration; and
establishing (208), via a robotic agent controlled by one or more hardware processors, a safe corridor traversal for the modified navigation path through a confined corridor using a safe corridor traversal technique, for achieving decentralized safe navigation within a dynamic environment.
2. The method as claimed in claim 1, wherein the safe radius is assigned to each of the plurality of robotic agents, and
wherein the safe radius is dynamically adjusted based on the current position of the robotic agent and the current positions of the plurality of robotic agents within the sensing field of the robotic agent.
3. The method as claimed in claim 1, wherein the optimal control action is one of (i) a symmetric and (ii) non-symmetric,
wherein the symmetric optimal control action generated allows the robotic agent to obtain a symmetrical path, and
wherein the non-symmetric optimal control action generated allows the robotic agent to break the symmetrical path when the static or dynamic obstacle is in the symmetrical path.

4. The method as claimed in claim 1, wherein the obtaining the safe corridor traversal comprises iteratively performing a sequence of steps until the robotic agent attains the goal position, the sequence of steps comprising:
dividing the goal position of each of the plurality of robotic agents into a plurality of sub-goal regions;
defining, one or more elliptical regions for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions;
setting, the goal position of the robotic agent to a centroid of the intersection of first and second elliptical regions; and
achieving, the goal position by the robotic agent using the modified navigation path;
shifting, the goal position of the robotic to the centroid of the second and third elliptical regions, till each of the plurality of robotic agents reaches the goal positions allocated.

5. A robotic agent (100), comprising:
a memory (102) storing instructions;
one or more communication interfaces (106); and
one or more hardware processors (104) coupled to the memory (102) via the one or more communication interfaces (106), wherein the one or more hardware processors (104) are configured by the instructions to:
collect, real-time onboard sensor data acquired by each of a plurality of robotic agents in vicinity of the robotic agent;
determine a current position and a velocity of each of the plurality of robotic agents within the sensing field of the robotic agent by processing the real-time on-board sensor data;
iteratively perform a sequence of steps to determine a navigation path towards a goal position allocated to the robotic agent, wherein the sequence of steps, executed in context of the current position and the velocity of each of the plurality of robotic agents, and an initial position of the robotic agent, the plurality of steps comprising:
identifying a plurality of static obstacles and a plurality of dynamic obstacles within the sensing field of the robotic agent based on the goal position and a safe radius associated with the robotic agent;
modelling an optimization problem using a Control Lyapunov Fields (CLF-QP) technique based on the current position of the robotic agent, the current position of each of plurality of robotic agents for a current time instant, the safe radius and the goal position;
modelling a set of non-convex elements in the optimization problem and a set of collision avoidance constraint using a Sequential Convexification Programming (SCP), wherein SCP allows convexification of the optimization problem;
adding a random zero mean Gaussian noise to the current position of the robotic agent, the current position of each of the plurality of robotic agents;
generating, an optimal control action at the time instant for of the robotic agent in a decentralized manner by solving the optimization problem;
executing, the optimal control action at the current time instant by the robotic agent to attain an updated position at a next time instant; and
modifying, the navigation path of the robotic agent dynamically based on the optimal control action considering the safe radius, wherein the robotic agent navigates towards the goal position with each iteration; and
establish, a safe corridor traversal for the modified navigation path through a confined corridor using a safe corridor traversal technique, for achieving decentralized safe navigation within a dynamic environment.

6. The system as claimed in claim 5, wherein the safe radius is assigned to each of the plurality of robotic agents, and
wherein the safe radius is dynamically adjusted based on the current position of the robotic agent and the current positions of the plurality of robotic agents within the sensing field of the robotic agent.
7. The system as claimed in claim 5, wherein the optimal control action is one of (i) a symmetric and (ii) non-symmetric,
wherein the symmetric optimal control action generated allows the robotic agent to obtain a symmetrical path, and
wherein the non-symmetric optimal control action generated allows the robotic agent to break the symmetrical path when the static or dynamic obstacle is in the symmetrical path.

8. The system as claimed in claim 5, wherein the robotic agent is configured to obtain the safe corridor traversal by iteratively performing a sequence of steps until the robotic agent attains the goal position, the sequence of steps comprising:
dividing the goal position of each of the plurality of robotic agents into a plurality of sub-goal regions;
defining, one or more elliptical regions for each of the plurality of sub-goal regions wherein, an intersection of each of the plurality of sub-goal regions is represented by an intersection of the one or more elliptical regions;
setting, the goal position of the robotic agent to a centroid of the intersection of first and second elliptical regions; and
achieving, the goal position by the robotic agent using the modified navigation path;
shifting, the goal position of the robotic to the centroid of the second and third elliptical regions, till each of the plurality of robotic agents reaches the goal positions allocated.