Modular self-reconfigurable aerial robot self-reconfiguration method based on controllable margin

Abstract

The application discloses a self-reconfigurable method of a modular self-reconfigurable aerial robot based on controllable margin. When a unit body fails, self-reconfiguration is performed according to the following method: the controllable margin of an assembly body formed by the fault unit body in different optional positions is calculated, and the position of the fault unit body that makes the controllable margin of the assembly body formed by the fault unit body maximum is selected as the optimal position of the fault unit body, and then the optimal assembly body structure is obtained; the optimal assembly body structure is taken as a self-reconfiguration target, and the controllable margin of a basic unit during disassembly and assembly in a self-reconfiguration process is greater than a controllable threshold as a constraint condition, and self-reconfiguration of the modular self-reconfigurable aerial robot is performed; the controllable margin refers to the minimum value of force generated by the system in all directions after interference is offset. The application takes a quantifiable controllability index as a guide, guarantees controllability in a reconfiguration process, effectively reduces time complexity of reconfiguration, and improves safety and reliability of the reconfiguration process.

Description

Technical Field

[0001] This invention relates to a modular self-reconfigurable aerial robot self-reconfiguration method. Background Technology

[0002] Modular self-reconfigurable aerial robots (MSRARs) use flying units that can move freely in three-dimensional space as modules, providing rapid response in time-sensitive situations. These flexible modules can navigate autonomously quickly in cluttered and challenging environments with obstacles. They can then rendezvous where tasks need to be performed. However, unlike ground-based collaboration, aerial collaboration is more complex due to flight uncertainties and safety concerns, and flying modules are prone to failure. For multi-robot systems, this can be due to component failures (e.g., sensor malfunctions) or physical failures caused by energy depletion (e.g., battery exhaustion). Furthermore, self-reconfiguration is a complex task requiring intricate planning and coordination between modules to move and reach their target locations.

[0003] To reduce the mutual impact of inter-module failures, recent research has proposed a self-reconfiguration technique for MSRAR (Micro-Range Array) to adjust its configuration based on propeller failures, enabling continued mission execution while efficiently utilizing resources and mitigating the effects of rotor failures. They transformed the problem into a mixed-integer linear programming (MILP) problem to find the optimal combination, aiming to reduce the impact of failures on the aircraft's trajectory and minimize disassembly and reassembly steps, thus solving the self-reconfiguration problem after a failure. However, this algorithm, based on optimization in a simulation environment, has limitations. It requires pre-calculation of the assembly tree of the failure sequence offline to view the configuration tree when a failure occurs at a specific location. Furthermore, this method ignores certain uncontrollable factors in the combination process, and due to limitations in the objective function and the accuracy of the simulation's physics engine, uncontrollable assemblies may arise during the reconfiguration process. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a modular self-reconfiguration method for aerial robots based on controllable margin. With quantifiable controllability indicators as guidance, the method can effectively reduce the time complexity of the reconfiguration process and improve the safety and reliability of the reconfiguration process while ensuring the controllability of the reconfiguration process.

[0005] The present invention specifically adopts the following technical solutions to solve the above-mentioned technical problems:

[0006] A self-reconfiguration method for a modular self-reconfigurable aerial robot based on controllable margin is disclosed. The modular self-reconfigurable aerial robot is an assembly composed of multiple interconnected unit cells, wherein each unit cell is a rotorcraft. When a unit cell fails, self-reconfiguration is performed as follows: the controllable margin of the assembly formed by the failed unit cell in different selectable positions is calculated, and the position of the failed unit cell that maximizes the controllable margin of the assembled assembly is selected as the optimal position of the failed unit cell, thereby obtaining the optimal assembly structure. The optimal assembly structure is used as the self-reconfiguration target, and the controllable margin of the basic units disassembled and assembled during the self-reconfiguration process is greater than a controllable threshold as a constraint condition for self-reconfiguration of the modular self-reconfigurable aerial robot. The controllable margin refers to the minimum value of the forces generated by the system in all directions after canceling out disturbances.

[0007] Preferably, for each faulty unit, the controllability margin of all possible assemblies that the faulty unit and its adjacent units can form is calculated, and the assembly with the fewest number of units is found from the assemblies with a controllability margin greater than a preset controllable threshold as the minimum fault controllable assembly of the faulty unit; during the self-reconfiguration process, the normal unit and the minimum fault controllable assembly of each faulty unit are used as the basic units for disassembly and assembly.

[0008] More preferably, the controllable threshold is 0.

[0009] Preferably, when calculating the controllability margin of the assembly formed by the faulty unit in different optional positions, for symmetrical combination forms, only the controllability margin of one of them is calculated.

[0010] Preferably, during the self-reconfiguration process, the basic unit that minimizes the reduction in the controllability margin of the assembly is disassembled first, and the basic unit that maximizes the improvement in the controllability margin of the assembly is assembled first.

[0011] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0012] This invention provides a quantifiable controllability criterion for multi-unit systems based on positive controllability theory. Compared with existing methods that rely on controllability criterion matrices, this invention quantifies the controllability criterion and solves the problems of existing technologies requiring a known control efficiency matrix and depending on the control allocation method. This invention optimizes the disassembly and assembly sequence by directly calculating the optimal fault assembly and the minimum fault controllable assembly, which can effectively reduce the time complexity of self-reconfiguration and solve the shortcomings of existing methods that ignore certain combined uncontrollable factors. Attached Figure Description

[0013] Figure 1 Define schematic diagrams for unit cells and assemblies;

[0014] Figure 2 This is a schematic diagram of the self-reconfiguration process of the present invention;

[0015] Figure 3 A schematic diagram illustrating the reconstruction process and controllability of a single unit failure in a 3x2 assembly;

[0016] Figure 4 This diagram illustrates the reconstruction process and controllability of a single unit failure in a 3x3 assembly. Detailed Implementation

[0017] To address the shortcomings of existing technologies, this invention provides quantifiable controllability criteria for multi-unit systems based on positive controllability theory. Compared with existing methods that rely on controllability criteria matrices, this invention quantifies the controllability criteria, solving the problems of existing technologies requiring known control efficiency matrices and depending on control allocation methods. By directly calculating the optimal fault assembly and the minimum fault controllable assembly, the disassembly and assembly sequence is optimized, effectively reducing the time complexity of self-reconfiguration and addressing the shortcomings of existing methods that ignore certain combined uncontrollable factors.

[0018] The present invention specifically adopts the following technical solutions to solve the above-mentioned technical problems:

[0019] A self-reconfiguration method for a modular self-reconfigurable aerial robot based on controllable margin is disclosed. The modular self-reconfigurable aerial robot is an assembly composed of multiple interconnected unit cells, wherein each unit cell is a rotorcraft. When a unit cell fails, self-reconfiguration is performed as follows: the controllable margin of the assembly formed by the failed unit cell in different selectable positions is calculated, and the position of the failed unit cell that maximizes the controllable margin of the assembled assembly is selected as the optimal position of the failed unit cell, thereby obtaining the optimal assembly structure. The optimal assembly structure is used as the self-reconfiguration target, and the controllable margin of the basic units disassembled and assembled during the self-reconfiguration process is greater than a controllable threshold as a constraint condition for self-reconfiguration of the modular self-reconfigurable aerial robot. The controllable margin refers to the minimum value of the forces generated by the system in all directions after canceling out disturbances.

[0020] Preferably, for each faulty unit, the controllability margin of all possible assemblies that the faulty unit and its adjacent units can form is calculated, and the assembly with the fewest number of units is found from the assemblies with a controllability margin greater than a preset controllable threshold as the minimum fault controllable assembly of the faulty unit; during the self-reconfiguration process, the normal unit and the minimum fault controllable assembly of each faulty unit are used as the basic units for disassembly and assembly.

[0021] More preferably, the controllable threshold is 0.

[0022] Preferably, when calculating the controllability margin of the assembly formed by the faulty unit in different optional positions, for symmetrical combination forms, only the controllability margin of one of them is calculated.

[0023] Preferably, during the self-reconfiguration process, the basic unit that minimizes the reduction in the controllability margin of the assembly is disassembled first, and the basic unit that maximizes the improvement in the controllability margin of the assembly is assembled first.

[0024] To facilitate public understanding, the technical solution of the present invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings:

[0025] In this embodiment, the modular self-reconfigurable aerial robot's assembly and unit are as follows: Figure 1 As shown, the unit has a cuboid shape and is composed of lightweight carbon fiber rods and a quadcopter drone; the unit can be connected to each other on a horizontal plane to form an assembly.

[0026] The total thrust and roll, pitch and yaw moments of the assembly are represented by F, M, and M, respectively. x M y M z This indicates that the control efficiency is calculated based on the propeller lift results in all unit cells using the control efficiency matrix:

[0027]

[0028] Where, η i,j ∈[0,1] (j=1,...,4) is used to characterize the damage or failure coefficient of thruster j in the i-th unit. If it is a complete failure, then η i,j =0; if perfectly healthy, then η i,j =1. d is the distance from the thruster to the center of the unit, C T and C M These are the thrust and pull coefficients of the propeller, respectively. i,j Let be the thrust of thruster j in the i-th unit. In fact, the dynamic model of a multi-rotor UAV is nonlinear, exhibiting certain aerodynamic damping and stiffness. However, if the multi-rotor helicopter is hovering, aerodynamic damping and stiffness can be ignored. The linear dynamic model of MSRAR in hovering state is as follows:

[0029]

[0030] in,

[0031]

[0032] in, This represents the position of the MSRAR in the Z direction, where φ, θ, and ψ are the attitude angles. The velocity in the Z direction, ωX represents the angular velocity in the X, Y, and Z directions, respectively. n is the number of unit cells in the assembly, m is the mass of each unit cell, g is the gravitational acceleration, I is the identity matrix, and 0 is the zero matrix. Since the unit cell is an approximately centrosymmetric object, J... xy =J xz =J yz =0. The moment of inertia J of the assembly. A,f It can be based on the rotational inertia J of the unit body u Using the parallel axis theorem, we can calculate:

[0033]

[0034] Where, x i ,y i This refers to the position of the unit cell in the assembly coordinate system. In practice, each propeller can only provide thrust in one direction (vertically upwards or downwards, perpendicular to the fuselage), i.e., f. j ∈[0,T j (j=1,...,n) r Therefore, the thrust vector of the propeller... Subject to the following constraints:

[0035]

[0036] in, Based on the geometric layout of the unit-type aircraft, the rotor thrust f i (j=1,...,n r ) and the total tension and torque of the system u f The mapping relationship between them is

[0037] u f =B f f

[0038] Among them, matrix It is the control efficiency matrix. Based on the control efficiency matrix, MSRAR's B... f The parameterized representation is as follows:

[0039] B f =[B f,1 B f,2 … B f,i ]

[0040] This invention proposes a controllability margin (CM) calculation method for MSRAR systems to quantify the controllability of MSRARs composed of multiple units. First, based on the propeller thrust vector constraint and the mapping relationship between total thrust / torque, u is obtained. f The constraints are as follows:

[0041] Ω={u f |uf =B f f, f∈U f}

[0042] Where, μ f This refers to the range of the tension vector. The constraint set for the control variable u is as follows:

[0043] U = {u|u = u} f -g,u f ∈Ω}

[0044] The CM metric will be defined below:

[0045]

[0046] in, Ω is the boundary, Ω C It is the complement of Ω. Assume... have That is, α is not an interior point of Ω; assuming Let α be an interior point of Ω. According to the above equation, assuming g is an interior point of Ω, we have...

[0047]

[0048] Among them, u' f = [u1 u2 / d u3 / d u4 / d], where u1, u2, u3, and u4 are u f The first to fourth terms, d represents the lever arm in the corresponding direction. It is the minimum value after the forces generated by the MSRAR system in all directions have canceled out interference. (Indicator) This can be viewed as the CM index of the system, serving as a basis for the system's controllability during failures. Here, CM is the MSRAR system vector u. f The remaining margin after offsetting g. It can be proven that the necessary and sufficient condition for the MSRAR system to be controllable is...

[0049] Multi-robot systems such as MSRAR allow for interconnection and expansion between unit modules, thereby increasing the number of faults that can be handled. Based on the aforementioned multi-unit CM calculation method, this invention proposes a self-reconfiguration method suitable for multi-unit systems. This method assumes that each unit module contains important equipment or data information that cannot be discarded, and that self-reconfiguration can be completed without landing during mission execution. Therefore, the assembly can reassemble in mid-air after a fault, improving residual fault tolerance.

[0050] After a failure or performance degradation occurs, the observer reports the failure or degradation status of each unit's propeller. After obtaining the performance status of all units, Algorithm 1 is first used to calculate the optimal combination after the failure, ensuring that the new combination has higher controllability, that is, ensuring that the assembly still has fault tolerance after the next unknown failure occurs; then Algorithm 2 is used to calculate the minimum controllable assembly after the failure of the unit and its neighboring units, so that the failure unit can be safely transferred.

[0051] The specific process for calculating the optimal faulty assembly is as follows:

[0052] 1) Calculating the possible positions of the faulty unit: Since faults in symmetrical positions have the same impact on the assembly, it is equivalent to performing rotation and symmetry operations on the assembly. To reduce computational complexity, this invention first calculates the symmetry of all combinations, removes symmetrical configurations, and then calculates the possible positions of the faulty unit;

[0053] 2) Calculate the optimal position of the faulty unit: Since the assembly and the unit use the same coordinate system, in order to facilitate the control of the reconstructed assembly, this invention performs CM analysis on the optional positions of the faulty unit in 1) without rotating the unit, and finds the combination form with the maximum CM.

[0054]

[0055] Where P represents the position information of the unit and E represents the assembly method.

[0056] After completing the optimal assembly calculation, a self-reconfiguration process is required. However, faulty units are often uncontrollable and need to be combined with nearby units to achieve joint flight, such as... Figure 2 As shown. Current optimization methods simply connect a normal unit (target assembly 2) next to the faulty unit, which cannot achieve controllable flight. In contrast, this invention uses Algorithm 2 to calculate the minimum controllable fault assembly formed by the faulty unit and its neighboring units, thus ensuring the safe and reliable transfer of the faulty unit from a controllability theory perspective.

[0057] The specific steps for calculating the minimum fault-controllable assembly are as follows:

[0058] 1) Locate the unit cell adjacent to the faulty unit cell: Based on the serial number coding principle, locate the unit cell adjacent to the faulty unit cell;

[0059] 2) Find the minimum number of controllable fault assemblies: Perform CM calculations on the faulty unit and its neighboring units, and determine if the CM is greater than a set threshold (usually 0). If it is less than the threshold, continue to increase the number of neighboring units until the remaining controllability is greater than the threshold. At the same time, under the same number of units, select the combination with the highest controllability.

[0060]

[0061] Through the calculations of Algorithm 1 and Algorithm 2, the optimal configuration of the assembly and the calculation method of the smallest fault-controllable assembly can be obtained, ensuring the safe and controllable transfer of faulty unit bodies and the further fault tolerance capability of the new combination form.

[0062] Based on the above calculations of the optimal fault assembly and the minimum fault controllable assembly, a simplest disassembly and assembly sequence can be determined, enabling the faulty unit to be safely transported to the optimal position and simultaneously forming the next assembly.

[0063] The specific steps for disassembling and assembling this invention are as follows:

[0064] 1) Calculate the relative offset of the fault unit: Calculate the relative offset of the current position of each fault unit and its position in the optimal fault assembly in the assembly coordinate system;

[0065] 2) Determine the spatial position of the optimal fault assembly: Fix the fault unit with the smallest relative offset, take it as the position of the optimal assembly in the world coordinate system, set its relative offset to 0, and add the relative offsets of the remaining fault units to the smallest relative offset.

[0066] 3) Calculate the transfer order of faulty units: Calculate the increase value of the system CM after each faulty unit is transferred to the corresponding position in the optimal assembly, and determine the transfer order of faulty units according to the decreasing order of CM values;

[0067] 4) Locate the unit adjacent to the faulty unit in this transfer: Based on the assembly's serial number coding principle, locate the unit adjacent to the faulty unit;

[0068] 5) Find the minimum number of controllable assemblies for the faulty unit in this transfer: Perform CM calculation on the faulty unit and its neighboring units, and determine whether the CM is greater than the set threshold. If it is less than the threshold, continue to increase the number of neighboring units until the remaining controllability is greater than the threshold. At the same time, under the same number of units, select the combination with the highest controllability;

[0069] 6) Transferring Faulty Units: Each faulty unit requiring transfer is transferred according to the transfer order using the minimum controllable assembly constructed in step 7). During disassembly and assembly, multiple options may be available; prioritize steps that minimize CM during disassembly and maximize CM during assembly.

[0070] 7) Controllability Guarantee of Remaining Assemblies: Ensure that after the smallest controllable assembly is detached from the assembly during the transfer of the faulty unit, the CM value of the assembly composed of the remaining units is greater than the threshold. If it is less than the threshold, the normal units in the remaining assembly need to be disassembled and reassembled in advance according to the minimum number of moves and steps to improve its CM value;

[0071] 8) Check the optimal fault assembly: After all faulty units are moved to their corresponding positions in the optimal fault assembly, check whether the new assembly conforms to the target assembly. If not, disassemble the normal units and move them to their corresponding positions in the target assembly. During disassembly and assembly, there may be multiple options; prioritize steps that result in a smaller reduction in CM value during disassembly and a larger increase in CM value during assembly.

[0072] This invention selected one case study to conduct CM analysis of the self-reconstruction process. Figure 3 This invention describes a process for relocating a faulty unit (unit 1) in a 3x2 assembly to unit 2 via self-reconfiguration. When a fault occurs in step 1, the relative offset of the coordinate system is calculated using the faulty unit as the coordinate center of the optimal faulty assembly. Units not within the optimal faulty assembly's coordinate system are then reassembled into the optimal faulty assembly through disassembly (steps 2 and 4) and assembly (steps 3 and 5). Multiple choices may be made during disassembly and assembly; steps with minimal impact on CM (component stability) during disassembly (step 2, solid line) and significant improvement in CM during assembly (step 3, solid line) are prioritized. Benefiting from the quantifiable CM analysis proposed in this invention, disassembly and assembly steps with greater CM can be easily selected. Compared to opposing disassembly and assembly methods, the average CM improvement is 9.06%, further enhancing the safety and reliability of the self-reconfiguration process.

[0073] Taking a 3x3 assembly as an example, this invention analyzes the controllability of the comparison method and the method proposed in this invention during the disassembly and assembly process. Figure 4 The self-reconfiguration process of the reference method after a single unit failure is shown. The comparative method uses MILP to find the optimal assembly, with the objective function being to maximize the torque generated by the weighted average of the trajectory-based x and y axes. This method ignores the controllability of the assembly with the failed unit. Due to the constraints considered and the accuracy of the simulation engine, uncontrollable (or degraded to controllable) assemblies are generated during disassembly. Specifically, Figure 4During the 5th and 6th disassembly steps, the remaining controllability of the assembly containing the faulty unit was -0.2824, which is less than 0 and therefore uncontrollable. The method of this invention has a higher controllability margin (CM) throughout the process, with an average controllability improvement of 177.16%, meaning higher flight safety controllability.

[0074] Figure 4 The paper also demonstrates the relationship between the number of assembly steps and the minimum CM during disassembly and assembly when the unit at position 8 in a 3x3 assembly fails, comparing the present invention and the reference method. Both methods found the same optimal fault-optimal assembly (steps 7 and 12). The difference lies in the fact that, after a fault occurs (step 1), the comparison method requires 11 steps (2-12), while the present invention only requires 6 steps (2-7) to reconstruct the optimal fault assembly, reducing the time spent on MSAR self-reconstruction and minimizing the impact of the faulty unit on the overall system.

Claims

1. A modular self-reconfigurable aerial robot self-reconfiguration method based on controllable margin, wherein the modular self-reconfigurable aerial robot is an assembly composed of multiple interconnected unit bodies, wherein each unit body is a rotorcraft; characterized in that, When a unit failure occurs, self-reconfiguration is performed as follows: The controllability margin of the assembly formed by the failed unit at different selectable positions is calculated, and the position of the failed unit that maximizes the controllability margin of the assembled assembly is selected as the optimal position of the failed unit, thus obtaining the optimal assembly structure. Using the optimal assembly structure as the self-reconfiguration target, and with the controllability margin of the basic units disassembled and assembled during the self-reconfiguration process exceeding a controllable threshold as a constraint, the modular self-reconfigurable aerial robot is self-reconfigured. The controllability margin refers to the minimum value of the forces generated by the system in all directions after canceling out disturbances. The controllability margin is calculated as follows: First, based on the propeller's thrust vector constraint and the mapping relationship between total thrust / torque, the torque is obtained. The constraints are as follows: in, It is the range of the tension vector f, B f It is the control efficiency matrix; Control quantity The constraint set is as follows: g is the acceleration due to gravity; The controllability margin index will be defined below: in, for boundary, yes complement; assumption ,have ,Right now No An interior point; assume , for An interior point; Based on the above formula, assume yes Interior point, has in, , They are respectively The first to fourth terms, d represents the lever arm in the corresponding direction.

2. The modular self-reconfiguration method for aerial robots based on controllable margin as described in claim 1, characterized in that, For each faulty unit, calculate the controllability margin of all possible assemblies that the faulty unit and its adjacent units can form, and find the assembly with the fewest number of units among the assemblies whose controllability margin is greater than a preset controllable threshold as the minimum fault controllable assembly of the faulty unit. During the self-reconfiguration process, the smallest fault-controllable assembly of normal unit cells and each faulty unit cell is used as the basic unit for disassembly and assembly.

3. The modular self-reconfiguration method for aerial robots based on controllable margin as described in claim 2, characterized in that, The controllable threshold is 0.

4. The modular self-reconfiguration method for aerial robots based on controllable margin as described in claim 1, characterized in that, When calculating the controllability margin of the assembly formed by the faulty unit in different optional positions, for symmetrical combination forms, only the controllability margin of one of them is calculated.

5. The modular self-reconfiguration method for aerial robots based on controllable margin as described in claim 1, characterized in that, During the self-reconfiguration process, the basic unit that minimizes the controllability margin of the assembly is disassembled first, and the basic unit that maximizes the controllability margin of the assembly is assembled first.

Images