9 results about "Iterative learning algorithm" patented technology

The invention relates to the technical field of pulse power and electromagnetic manufacturing, and discloses a pulse magnetizer which comprises a charging module, a capacitance matrix module, an acquisition module and a control module. The acquisition module adopts a Kelvin four-wire system structure to synchronously acquire a voltage and current discrete sequence at a discharge moment; the control module is based on a heterogeneous architecture of a digital signal processor and a field programmable gate array, equivalent resistance and inductance parameters of a load are identified in real time by using a regularization least square algorithm, and a closed-loop control logic based on single waveform reverse analysis is adopted by a system core: a nonlinear equation is solved by using a Newton iteration method to reconstruct a capacitance matrix; the pulse width is accurately locked; and meanwhile, an iterative learning algorithm is combined to correct the charging voltage, compensate the capacitance discrete quantization deviation and lock the peak current. According to the invention, the self-adaptive compensation of the magnetizing waveform to the load parameter drift is realized, the current repetition precision is obviously improved, and the service life of the system is obviously prolonged.

The invention discloses a piezoelectric actuator compensation method based on an LSTM inverse model and iterative learning, and relates to a piezoelectric actuator compensation method. The invention aims to solve the problems of obvious nonlinearity and time-varying characteristics during frequency control speed output of a piezoelectric actuator. The method comprises the steps of 1, data acquisition and LSTM inverse model training; firstly, a controller program is written, then data is cleaned, abnormal values are eliminated, samples are recombined, finally, a double-layer LSTM network is constructed, and a model is exported after a training set is trained through a Huber loss function; step 2, iteratively learning a control flow; the iterative learning process is used as an independent medium-low priority task, is triggered after one operation cycle is ended, and reads a speed tracking error sequence in the whole operation cycle after the operation of the LSTM inverse model is ended from the shared memory to obtain data; and step 3, designing an LSTM-based inverse model and an iterative learning algorithm. The invention belongs to the technical field of precision motion systems.

The invention relates to the field of unmanned aerial vehicle control, in particular to an unmanned aerial vehicle predefined time fault-tolerant control method based on an iterative learning neural network, and the method comprises the following steps: S1, building an outer ring position system model and an inner ring attitude system model of a three-rotor unmanned aerial vehicle under external interference and actuator faults; s2, designing a neural network structure to realize disturbance real-time estimation compensation, and updating a neural network weight matrix through an iterative learning algorithm; and S3, respectively designing an outer ring position pre-defined time fault-tolerant controller and an inner ring attitude pre-defined time fault-tolerant controller by adopting an inner and outer ring tracking control strategy, estimating and compensating an actuator fault online by adopting a self-adaptive algorithm, and ensuring that the unmanned aerial vehicle tracks quickly and stably while the system pre-defined time is converged. According to the invention, through collaborative design of neural network compensation and the predefined time adaptive fault-tolerant controller, the influence of the actuator fault on the system stability is reduced, and the control performance of the unmanned aerial vehicle during external disturbance is improved.

The application discloses a trajectory tracking control method for an unmanned system without an initial stabilizing learning strategy, relates to the technical field of unmanned systems, and can overcome the dependence of an initial stabilizing control strategy on a system model and ensure that an unmanned vehicle system realizes trajectory tracking of a leader under a designed reinforcement learning controller. First, an unmanned system model is established; a trajectory tracking controller of a follower and a corresponding cost function are designed to realize optimal trajectory tracking of the follower; an offset factor is introduced on the basis of an augmented system of the unmanned system to construct a new closed-loop system; system data are generated according to the new closed-loop system, and the system data are collected; and a policy iteration learning algorithm is used to find an optimal control strategy to realize model-free trajectory tracking control without dependence on an initial stabilizing control strategy.

The application discloses a kind of LVDT sensor measurement precision compensation method based on key component position adjustment, the position of lever fulcrum is optimized by finite element analysis, the mechanical model containing temperature variable is constructed, the influence of the thermal expansion coefficient of sensor material at different working temperatures on the mechanical properties of lever is considered, to ensure that the lever is uniformly stressed within the full range, and maintain the optimal stress state within the full temperature range, reduce the core movement deviation and nonlinear error caused by lever structure and temperature change, improve the measurement accuracy of sensor. Iterative learning algorithm is used to compensate the measurement data, based on multiple measurement data, the residual is corrected in reverse, the compensation parameters are corrected in reverse, and the momentum term is introduced to accelerate the algorithm convergence, and the momentum coefficient is adaptively adjusted according to the dynamic response characteristics of the sensor. This dynamic adjustment compensation method can better adapt to the measurement error change under different working conditions, greatly improves the compensation effect.

The invention provides a multi-point sensor and algorithm-optimized absorption tower slurry density balance regulation and control method. The method comprises the following steps: mapping a low-density slurry region at the upper part of an absorption tower according to a dynamic change trend of density difference, and processing the region by adopting a grouping clustering algorithm to obtain a clustering result of spatial distribution change; according to the clustering result of the spatial distribution change and the corrosion environment influence compensation factor in the absorption tower, whether the factor reduces the monitoring precision or not is judged, and an adjusted slurry density measurement value sequence is obtained; extracting a process stability index from the improved desulfurization process efficiency monitoring data, and processing the index by adopting a dynamic adjustment algorithm to obtain an optimized slurry density equilibrium state; and updating deployment parameters of a multi-point sensor array according to the optimized slurry density equilibrium state, and processing the parameters by adopting an iterative learning algorithm to obtain enhanced real-time monitoring precision.

The invention discloses a track correction and control method for a multi-axis electro-hydraulic servo system of a radial forging machine, and provides a cooperative control scheme fusing an extended state observer and online iterative learning for the composite control problem of coexistence of random high-frequency forging impact and workpiece axial plastic extension accumulated deviation in the radial forging process. According to the method, by establishing and discretizing a system model, on one hand, real-time estimation and feed-forward compensation are carried out on total disturbance including random impact by using an extended state observer; on the other hand, multi-step state prediction is carried out based on the discrete model, an online iterative learning algorithm is adopted to carry out iterative correction on prediction errors caused by slow change factors such as workpiece extension, and track correction is generated. Finally, disturbance compensation and trajectory correction jointly act on a controller, synchronous suppression and compensation of high-frequency random interference and low-frequency trend deviation are achieved, and trajectory tracking precision and robustness of the multi-axis system under strong impact and nonlinear working conditions are remarkably improved.

The present application relates to the technical field of pulse power and electromagnetic manufacturing, and discloses a pulse magnetizing machine, which comprises a charging module, a capacitor matrix module, an acquisition module and a control module; the acquisition module synchronously acquires voltage and current discrete sequences at the discharge moment in a Kelvin four-wire structure; the control module is based on a heterogeneous architecture of a digital signal processor and a field programmable gate array, uses a regularized least square algorithm to identify equivalent resistance and inductance parameters of a load in real time, and adopts a closed-loop control logic based on single waveform reverse analysis as a system core; a Newton iteration method is used to solve nonlinear equations to reconstruct the capacitor matrix and accurately lock the pulse width; at the same time, the charging voltage is corrected by combining an iterative learning algorithm, the capacitor discrete quantization deviation is compensated, and the peak current is locked. The present application realizes adaptive compensation of the magnetizing waveform to load parameter drift, and significantly improves the current repetition accuracy and the operation life of the system.