In the study of 3-Dimensional (3D) local feature descriptor, it is difficult to strike a balance among accuracy, matching time and memory consumption. To solve this problem, an adaptive binary simplification method for 3D feature descriptor was proposed based on the standard deviation principle in statistical theory. First, different binary feature descriptors were generated by changing the binarization unit length and the number of standard deviations in the simplification model, which were applied into the currently widely used Signature of Histogram of OrienTations (SHOT) descriptor, and the optimal combination of binarization unit length and the number of standard deviations was determined by experiments. Finally, the simplified descriptor under the optimal combination was named Standard Deviation feature descriptor for Signature of Histogram of OrienTations (SD-SHOT). Experimental results show that compared with the SHOT descriptor without simplification, SD-SHOT reduces the key point matching time to 1/15 times and the memory occupancy to 1/32 times of SHOT; compared with the existing mainstream simplification methods such as Binary Feature Descriptor for Signature of Histogram of OrienTations (B-SHOT), SD-SHOT has the optimal comprehensive performance. In addition, the validity of the proposed method is verified in the actual robot sorting scene consisting of five different categories of objects.

In order to improve the anti-noise performance of the traditional Compressed Sensing (CS) recovery algorithm, a kind of Adaptive Compressed Sensing (ACS) algorithm was proposed based on the idea of observation matrix optimization and adaptive observation. The observed energy was all allocated in the support position estimated by the traditional CS recovery algorithm, which could effectively improve the observed Signal-to-Noise Ratio (SNR) owing to the support positions contained in the estimated support set. Then, the optimal new observation vector was derived from the perspective of observation matrix optimization, that is, its nonzero part was designed as the eigenvector of Gram matrix. The simulation results show that, the energy growth rate of non-diagonal elements of Gram matrix is less than that of the traditional CS algorithm with the increase of the number of observations. And the reconstruction normalized mean square error of the proposed algorithm is respectively lower than that of the traditional CS algorithm and the typical Bayesian method above 10 dB and 5 dB under the same conditions of number of observations, sparsity and SNR. The analysis shows that the proposed adaptive observation mechanism can effectively improve the energy efficiency and anti-noise performance of the traditional CS recovery algorithm.

With regard to the weighted transition system, the linear time properties were proposed. Firstly, the weighted transition system above semiring K was defined, the concepts of the weighted linear time properties were given, the upper, the lower and the closure of weighted linear time properties were determined by the weighted function; secondly some familiar weighted linear time properties and their relationships were discussed; thirdly the weighted safety property was mainly studied, the weighted regular safety property was defined through weighted automaton and closure of weighted regular safety property; finally, the checking method of the weighted regular safety was built based on weighted finite automaton. The checking was follows. Together with semiring and formal series, the product system was built over weighted transition and weighted finite automaton, the model checking about weighted safety property of weighted transition was transferred to verify the invariance of the product system, the algorithm and complexity were given. Finally, an example shows the model checking of weighted regular safety property is reasonable and efficient. The example result shows the proposed method can verify the safety of the weighted system.