Research On Test Generation For I_DDT Testing Based On Particle Swarm Optimization And Ant Algorithm

With the rapid development of the microelectronic technique, the level of integrated and complex of VLSI is increasing continuously, it makes the great challenge to the present test technology and test instrument.In the early 1980's,people have put forward the method of measuring IDDQ test and CMOS circuit have good compatibility, it can test the error and physical fault that the voltage test method can not check. Now, it becomes a popular and important method to the CMOS digital integrated circuit.In the middle age of the 90s, people have raised the idea of IDDT Testing. It attempted to find some faults that can't check out with other methods by observing and analysis the transient current when the circuit inside status has changed. This method acted as the addition to the traditional test method, it has gradually attracted the attention and studying in the field of studying and industry.In IDDT Testing, the core problem is giving what kind of excitation to excite the fault in any term, and it can be measured by instrument. The problem that Auto Test Pattern Generator (ATPG) shall be solved is how to derive the test excitation to the tested circuit.IDDT Testing needs two or more vectors once. The test vector making is rather complex. It has important point to raise test efficiency to find efficient test vector generation arithmetic. This paper emphasizes to study IDDT Testing generation arithmetic. From the activation conditions of the stuck-open fault, I have given two arithmetic about IDDT Testing, Particle swarm optimization algorithm and Ant algorithm to make test vector, and using the waveform simulator to model test to check the validity of the arithmetic. In IDDT Testing, the particle swarm optimization algorithm defines that the fault what the particle is searching, the position of particle defines test vector. Ant algorithm defines that the fault is food what the ant is searching, searching path defines test vector.To the feasibility of the IDDT Testing, the two arithmetic generate vector is acted as excitation. With the waveform simulation, it can give stuck-open fault test. By modeling test to the waveform simulation of the ISCAS'89 part circuit, it can get good fault coverage. The disperse Particle swarm optimization algorithm has internal variation characteristic and the great overall searching ability. In circuit test generation, its speech is fast, but to the VLSI test, the generative test vector is not the best. Ant algorithm is powerful in searching, and its speech is fast, the arithmetic test generation vector by SPICE simulation experiments showed that its result is effective.