Automatic Verification System Research Articles

CLASS: a CAD system for automatic synthesis and verification of asynchronous finite state machines

This paper describes algorithms and techniques underlying a CAD system called CLASS (Cirrus Logic Asynchronous Synthesis System) for automatic synthesis and verification of control circuits based on Asynchronous Finite State Machine (AFSM) specifications. AFSM specifications are transformed into Signal Transition Graphs, and then to State Graphs. Newly developed hazard-free synthesis techniques from State Graphs will be described. An efficient two-level hierarchical verification technique based on State Graph Contraction and Dill's verifier is used to verify the logic implementations against the state graph.

Black and white in accuracy assessment of megavoltage images: the medical decision is often grey

Using different criteria for acceptance of the portal film taken at the first treatment session, a comparison was made of the relevance of the information obtained from such a single check for the subsequent irradiations. A total number of 234 verification films have been taken on 29 fields for 27 head and neck patients. Patients were immobilised with individual plastic masks fixed to the couch and treated on a 6-MV linac fitted with an automatic verification system. Field alignment was checked with a measurement in the anteroposterior (AP) and craniocaudal (CC) direction on each film. Referring to the simulated field, this group of patients was treated with excellent average precision (mean, −0.7 mm) and reasonable spread (s = 5 mm). The percentage of ‘large’ deviations (≥6 mm) occurring during the whole treatment course is proportional to the upper limit of deviation accepted in the first assessed field (for an upper limit <6 mm): it goes progressively up from 5% (AP-CC direction) to 17% (AP) and 13% (CC) for accepted magnitudes of deviation going from 2 mm to 6 mm in the first film. As the reproducibility of the different treatment series (s = 2 mm) is independent of the upper level of error accepted on the first film, this means that errors are mainly systematic errors coming from the transfer of the simulation unit to the treatment unit. Precision in a series of set-ups is always expressed by a Gaussian curve. A control port film at the first session is most frequently representative for the median value of the whole series and subsequent set-ups always show a spread of values. It is therefore important to choose an upper limit of the acceptable deviation on the first measurement which is small enough, as compared with the ‘unacceptable deviation’, to be sure that most of the subsequent set-ups can also be considered as correct. The acceptance limit of a single first image should therefore be directly linked to the known precision of a technique (determined by the standard deviation of the reproducibility distribution) and to the clinical situation involved. It should not be limited to a ‘black’ or ‘white’ decision with a single numerical value for different treatment modalities.

Automatic consistency verification of mechanical CAD files

Abstract A system for automatic verification of the exactness of the CAD files of a mechanical part with planar and cylindrical faces through 3D part reconstruction is described. A set of consistency checking rules are specified to facilitate the part reconstruction. The system not only reconstructs the 3D geometry from a correct drawing file, but also produces adequate error messages to aid the designer to make corrections. We shall present our method with illustrations. Experimental results indicate that the set of checking rules is very useful and the system is very effective.

In vivo dosimetry for tangential breast irradiation: role of the equipment in the accuracy of dose delivery

Entrance dose measurements (920) have been performed on both tangential treatment fields for 105 breast cancer patients. About half of them (52) were treated on a modern unit (Mevatron, Siemens), supplied with an automatic verification system, the other half (53) were treated on an old therapy unit (cobalt-60). A team of three radiographers worked on the modern unit, while on the old unit only one radiographer was responsible for the treatment set-up. A small systematic error ( + 1.4%) has been detected on the old unit: a small discrepancy existed in the region of the maximum dose between the measured and the published percentage depth-dose values (Br. J. Radiol., Suppl. 17), the latter being used for dose calculation. A striking difference has been observed for the rate of large deviations (deviations of 5% and more from the mean dose): 2.3% (10/430) for the modern unit versus 15% (75/490) for the old one. This shows clearly that the precision in dose delivery is strongly dependent on the characteristics and staffing of the therapy unit, the modern unit being supplied with a check-and-confirm system and with the beam accessory devices in a fixed position in the beam and a team of radiographers being responsible for the set-up.

Tangential breast irradiation: influence of technique of set-up on transfer errors and reproducibility

Using conventional portal films, the influence of the technique of set-up on the transfer error from simulator to treatment couch and on the subsequent reproducibility was made for the irradiated volume in the treatment of breast cancer. A total number of 376 portal films have been performed on 14 patients. All the patients were treated on a 6 MV Linac supplied with an automatic verification system excluding, however, the couch parameters. The overall precision of the treatment delivery is evaluated by the global analysis of the discrepancies between the simulator films and the different portal films. For the patient group lying on a inclined plane with (group 2) or without (group 1) fixed arm support, a narrow gaussian distribution is obtained in the anteroposterior (AP) direction with a SD of 4 mm. In the craniocaudal (CC) direction, the frequency of distribution of the patients treated without fixed arm support is much larger than in the other group: the SD is respectively 15.5 mm for the first and 5.5 mm for the second one. The reproducibility of the series of set-ups of the two groups estimated by reference to the mean value are similar in the AP direction. The SD within the series of portal films in the CC direction is 5.8 mm for the first group and 3.7 mm for the second group. Their comparison with the discrepancies of 15.5 mm and 5.5 mm when assessing the deviation between the simulated and portal films shows the importance of systematic errors. The localisation of the irradiation port can be improved by using support systems with fixed arm indicators.

Is it necessary to repeat quality control procedures for head and neck patients?

Using serial verification films for detection of localization errors and in vivo measurements of the delivered dose, a comparison was made of the information obtained from a single check on the first treatment session or from repeated checks in subsequent irradiations, leading to an assessment of the predictive value of a single check. A total number of 215 films and 261 entrance dose measurements have been performed on 34 fields for 10 head and neck patients. The patients are immobilized with individual plastic masks fixed on the couch and treated on a 6 MV linac, supplied with an automatic verification system excluding the couch parameters. The global results show Gaussian frequency distributions with standard deviations of 4 mm for port film measurements and 3.4% for the dose measurements. Large errors (> 5 mm displacement and > 4 % deviation from the expected dose) have been detected in 16% in the cranio-caudal direction and 24% in the antero-posterior direction with port films and in 15 % of the in vivo measurements. In order to identify the nature of the errors, which can be random or systematic, the first measurement is taken as the reference value and shows that consecutive measurements on the same field were reproducible with standard deviations of respectively 2.5 mm and 1.8 %. This means that a large part of the spread of the global results can be explained by systematic errors in the treatment preparation chain. With the first check, 6 out of 10 systematic localization errors and 7 out of 7 systematic errors leading to erroneous dose delivery have been detected. Therefore, most of the systematic errors, which affect the overall quality of the treatment, can be identified with the first check. The four systematic localization errors, missed with the first film, were of rather limited size: only one of them showed a mean displacement larger than 7 mm. Because the first measurement is an acceptable indication of the overall quality of the treatment delivery, the authors propose a code of practice for checking the treatment quality at the patient level.

Correctness verification of real-time programs

Two approaches to the correctness verification of real-time programs are presented. The first approach, static verification, corresponds to the correctness notion in systems programming, that is informally described by the software standard (ANSI/IEEE, 1983). The second one, dynamic correction, is a generalization of the exceptions handling idea (Anderson, 1981). The formal construction is based on the coupled process notion and the so-called local testability theorem. Automatic correctness verification system and a structure of dynamic correction module result from the formal investigations.

Automatic Verification of Sequential Circuits Using Temporal Logic

Verifying the correctness of sequential circuits has been an important problem for a long time. But lack of any formal and efficient method of verification has prevented the creation of practical design aids for this purpose. Since all the known techniques of simulation and prototype testing are time consuming and not very reliable, there is an acute need for such tools. In this paper we describe an automatic verification system for sequential circuits in which specifications are expressed in a propositional temporal logic. In contrast to most other mechanical verification systems, our system does not require any user assistance and is quite fast—experimental results show that state machines with several hundred states can be checked for correctness in a matter of seconds!

Two-level grammar as an implementable metalanguage for axiomatic semantics

Two-level grammars can define the syntax and the operational semantics of programming languages and these definitions are directly executable by interpretation. In this paper it is shown that axiomatic semantics can also be defined using a two-level grammar with the result being a partially automatic program verification system accomplished within the framework of a language definition. These results imply that a programming language can be defined operationally and axiomatically together in complementary definitions as advocated by Hoare and Lauer. Because two-level grammars are executable, these complementary definitions accomplish a system for interpreting and verifying programs.

Microcode verification using SDVS-the method and a case study

This paper describes SDVS (State Delta Verification System), its application to microcode verification, and the verification of a particular example referred to as the H-machine example. The example illustrates how particular microcode that interprets a computer instruction set can be proved correct and how this proof is accomplished with an existing, automated verification system.

Programming pearls: Writing correct programs

In the late 1960s people were talking about the promise of programs that verify the correctness of other programs. Unfortunately, it is now the middle of the 1980s, and, with precious few exceptions, there is still little more than talk about automated verification systems. Despite unrealized expectations, however, the research on program verification has given us something far more valuable than a black box that gobbles programs and flashes “good” or “bad”—we now have a fundamental understanding of computer programming. The purpose of this column is to show how that fundamental understanding can help programmers write correct programs. But before we get to the subject itself, we must keep it in perspective. Coding skill is just one small part of writing correct programs. The majority of the task is the subject of the three previous columns: problem definition, algorithm design, and data structure selection. If you perform those tasks well, then writing correct code is usually easy.

A New Automatic Logic Interconnection Verification System for VLSI Design

A new VLSI checking program, LIVES (Logic Interconnection VErification System), has been developed. LIVES verifies the geometrical layout data to determine whether or not it correctly reflects the original logic level description. An excellent LIVES feature, which no other programs have possessed yet, is that it can perform the check even when there are no identification marks on each logic gate in layout level description data. This feature is realized by employing new algorithms based on graph theory. In the verification procedure, LIVES Adopts graph isomorphism, which uses the new partitioning methods for vertices, especially tailored for the logic diagram. In the error detection procedure, a graph is divided into subgraphs, based on a new concept, "route-subgraph," and tests for graph isomorphism between subgraphs are performed. These algorithms enable precise and rapid search for any error points in the VLSI design. These new algorithms have been implemented and examined for practical feasibility. Experimental results have ensured that this program can detect errors in the VLSI layout patterns very quickly and precisely. LIVES will be a highly efficient tool in VLSI logic design by using it systematically combined with other CAD programs.

Verification of HDLC

A version of the standard high level data link control protocol has been verified as an experiment with an automated verification system. This paper tries to answer questions such as: How can HDLC and its properties be represented? To what extent can they be proved? What are the main obstacles in verifying protocols like HDLC.

A closer look at termination

Several methods for proving that computer programs terminate are presented and illustrated. The methods considered involve (a) using the "no-infinitely-descending-chain" property of well-founded sets (Floyd's approach), (b) bounding a counter associated with each loop (loop approach), (c) showing that some exit of each loop must be taken (exit approach), or (d) inducting on the structure of the data domain (Burstall's approach). We indicate the relative merit of each method for proving termination or non-termination as an integral part of an automatic verification system.

An Improved System for Automatic Speaker Verification

The interactive automatic speaker verification system described by Lummis [“Implementation of an On-Line Speaker Verification Scheme,” J. Acoust. Soc. Am. 52, 181(A) (1972)] has been augmented to include linear prediction parameters in addition to the already existing pitch and intensity analysis of sentence-long utterances. This improved system has been evaluated on a new and enlarged speaker population. A method for selecting optimum speaker-dependent features has been incorporated in this system which significantly improves its performance. The evaluation indicated that the verification error rate was less than 1% with respect to casual impostors and less than 4% with respect to well-trained mimics.

Speaker Verification

Speaker verification is defined as the process of confirming the purported identity of an informant. An automatic verification system is required to give a YES-NO decision with a corresponding measure of confidence. It would be desirable to begin with a set of ideal measurements that extract only the parameters in the voice relating to speaker identity. Since the significant parameters cannot be completely specified, an adaptive decision procedure has been utilized to focus on the characteristics that differentiate a particular speaker from a set of potential impostors. The adaptive linear-threshold element employed in these experiments is trainable for separating all input patterns into two classes. Such an approach is attractive because the measurement hardware is minimized, and the decision mechanism may be simulated quite readily on a digital computer. The present paper explores the extent to which adaptive linear-threshold elements are able to verify a person's identity based on quantized spectral patterns of his voice.

Automatic control and regulation of rotary chamotte-firing kilns

Automatic verification and control systems should be introduced in presently operating rotary kilns, and also those under construction. Work should be continued in order to establish ways and means of automatically checking the temperature in the firing zone, the quality of the finished chamotte, and the quality of the raw material, which will enable the complex automation of chamotte-firing rotary kilns to be effective.