55results about How to "Strong nonlinear fitting ability" patented technology

The invention discloses a three-dimensional dynamic grinding force detection device and its decoupling algorithm, which mainly solves the problems of low natural frequency, narrow measurable dynamic force range and piezoelectric grinding force measurement of the existing resistance strain grinding force measuring instrument. The instrument is complicated to manufacture, expensive, unable to detect static force, the multiple linear regression decoupling algorithm is highly dependent on the linearity of the detection device, and the decoupling effect is unstable. The three-dimensional dynamic grinding force detection device is mainly composed of a base, a column, a support, an elastic thin plate group, an upper cover plate and a resistance strain gauge group. The workpiece to be processed is installed on the upper surface platform through threaded connection. During grinding, the grinding force acts on the workpiece and is transmitted to the column through the upper surface platform, causing the column to deform accordingly, and then transmitted to the three sets of elastics installed in parallel with it. Thin plates, three groups of elastic thin plates are installed in two pairs vertically, causing the three groups of resistance strain gauges pasted on the elastic thin plates to produce corresponding resistance value changes, and after signal conditioning and collection, they are transmitted to the PC, realizing Three-dimensional dynamic grinding force detection. The decoupling algorithm uses the BP neural network as the decoupling model, and optimizes its initial weight and threshold parameters through the genetic algorithm to obtain the optimal decoupling model and realize the high-precision decoupling of the three-dimensional grinding force signal. The invention has ingenious structural design, high natural frequency, wide range of measurable dynamic force, and can measure static force at the same time, with low inter-directional coupling and simple manufacture, which greatly reduces the manufacturing difficulty and production cost of the three-dimensional dynamic grinding force detection device, and at the same time The decoupling algorithm has fast convergence, high precision and high reliability, and has good practical and popularization value.

The invention discloses a hen house environment ammonia gas concentration intelligent monitoring system which is composed of two parts of a hen house environment parameter acquisition platform and a hen house environment ammonia gas concentration intelligent prediction model. The hen house environment ammonia gas concentration intelligent monitoring system based on a CAN field bus is designed in view of difficulties that hen house environment ammonia gas concentration change has characteristics of non-linearity, high time lag, high inertia and time variation and hen house area is relatively large and ammonia gas concentration is difficult to accurately predict. The system is the intelligent monitoring system which realizes detection, intelligent prediction and management of the monitored hen house environment ammonia gas concentration by the hen house environment parameter acquisition platform and the hen house environment ammonia gas concentration intelligent prediction model, and the basis of performing high-quality and high-efficiency purification on hen house environment ammonia gas concentration is provided. The system has wide application prospect and great popularization value.

The invention relates to a method for achieving robot multi-axis-hole assembling through deep reinforcement learning, and belongs to the technical field of robot assembling. The method comprises the steps that in the training process, common experience data and expert experience data which are produced through a traditional fuzzy force control method and a deep reinforcement learning network on the basis of a simulation model are added into an experience data set, experience data are randomly drawn from the experience data set to train the deep reinforcement learning network, and therefore theassembling action of the network can rapidly achieve the assembling level of the traditional fuzzy control method and can exceed the assembling effect of the traditional fuzzy control method if training continues to be carried out. The deep reinforcement learning network trained by the simulation model is directly used for a practical robot multi-axis-hole assembling task. The experience data produced by the simulation model are used for training, the problem that practical assembling environment cannot provide enough training data is solved, and meanwhile the training cost is reduced.