abstract: Objective: In the project of RFID flip chip package’s development

the high speed manipulator have obvious nonlinear and time-variable characters. For this reason

the linear control can’t satisfy requirement

so applies nonlinear adaptive inverse control. The key of this method is to identify the high speed manipulator with three order Volterra nonlinear model in limited time and with sufficient accuracy. But it’s hardly to satisfy real-time requirement with conventional method. This paper proposes a quick identification algorithm to resolve the problem. Method: Firstly

constructs high order input vector by low order input vector according their structure character. Next

speeds up estimates high order kernels based on low order kernels according their correlation. finally

uses linear variable step-size LMS strategy in nonlinear algorithm

and proves it’s convergent using Lyapunov global stability theorem. Result: By the experiments of identified a manipulator using conventional method and proposed method respectively

the result shows this algorithm reduces identification time form 100 ms to 30 ms

improves convergent speed 3.3 times

and reduces maladjustment 93.3%

as well as have higher precision. Conclusion: It can satisfy the requirements of real-time and identification precision .