Biometric Technology In The Field Of Automotive Safety Applications

Based on the research situation of motor vehicle safety both in and aboard, thisarticle makes a deep discussion on motor vehicle safety, existing problems,enactments, study methods, etc. It also analyzes the active and passive safety level inthe world, and proposes a new way to solve the safety problem of the motor vehicle:an implementation of the biological identification technologies.Modern society witnesses a sharp development of network and informationtechnology. Because of the equity, privacy and protecting legal users, people have topass many kinds of ID identifications, such as financial trade, database access, staffmembers enrollment, private computer's accessing, and so on. The traditionalidentification techniques bases on password or Token-Ring, which have potentialhazards like loss, embezzlement, forgery, and stealing. Biological identificationtechnologies take advantage of a person's physical and behavior characteristics andconquer the drawbacks from the traditional identification. At present, ManyBiological identification technologies are used in the process of ID identifications,including iris identification technique, retina identification technique, countenanceidentification, ideograph identification, voice identification technique and fingerprintidentification system. As a quintessence of biological identification technologies,fingerprint identification has got the widest attention thanks to its uniqueness,un-changeability with the time and easy collection. And because of itsun-changeability and stability, fingerprint identification has a promising future, whichcould be widely used in the field of police, public security, confidential departmentand commerce.To meet the demand of practice, this paper makes a research on AutomaticFingerprint Identification System. After elaborates on the theories of fingerprintidentification, it also proposes a fingerprint identification system which combinedpassword with template self-adaptation and improved the safety and executive speedof the system. Fingerprint identification is a process that extracts the fingerprintfeatures and compares them with the existing templates in the system. To improve thespeed of the process, a classification based on the structure of fingerprint has beenused generally, so as to compare with the certain class. To a large system, however,the template numbers in each class are huge as well as the calculating works. As a result, it wastes time, which constrains the use of fingerprint identification in somecomprehensive systems. To solve these problems, this article proposes a new systemstructure for fingerprint classification, which could reduce fingerprint's comparingnumbers through this classification and increase its speed by a large margin. On theother side, those fingerprint feature templates saved in the system could changeadaptively according to the fingerprint's images inputted and approach user's realfingerprints step by step. In addition, based on the problem of bad qualitativefingerprint images obtained from actual sensor, this paper also proposes somemethods on image enhancement.In terms of fingerprint identification algorithm, integrate with the detailmatching and texture feature techniques, this paper introduces a novel fingerprintmatching algorithm, which mainly includes several aspects: fingerprint images'texture analysis and edge extraction bases on the wavelet theory; original matching offingerprint images bases on Genetic Algorithms; fingerprint spine's node matchingadopts an alterable delimitation box. We select two methods of point pattern to matchminutiae feature: the method bases on the images emendation and local vector matchmethod. The former adopts a changeable sized boundary box, which could avoid thedisadvantage of the fixed sized boundary box; the latter is introduced into somecharacteristic information, such as the line numbers and dispersion of direction anglebetween the center point and adjacent minutiae point, the direction angle of the centerpoint, etc. These improve the veracity of the minutiae feature matching.In the application of fingerprint identification, a new system combined withpassword and fingerprint is established in this thesis. PKI, based on the biologicalidentification technologies, a new simple but efficacious self-adaptive safetynumerical signature protocol is proposed by using fingerprint recognize, dynamicDES, RSA and verifiable secret sharing techniques, to achieve the function of fileencrypting, numerical certificate making, SSL security communication, encryptingand decrypting, and so on. A proof for its security is also given. It uses password asthe index of the fingerprint warehouse and finds the corresponding feature template,then validates the input fingerprint. This method could overcome the shortcomings,such as the forgery password and time-consuming in fingerprint identification. At thesame time, fingerprint's feature templates in the system could renew themselves, inorder to reflect the variation of the user's fingerprints, and consummate the registeredtemplates. This measure could affirm user's identity dynamically and ensure the reliability of data's transmissions. It solves problems not only the bidirectionalauthentication of the two communicated sides, but also the confidentiality andintegrality of the information (e.g. user's lived fingerprint's characteristics) during theprocess of communication through the network. It integrates biological identificationtechniques and cytological authentication protocol; the former takes the fingerprint'scharacteristics as the representation, the latter bases on the password system. Andprovides a more secure, reliable, facilitate and quick ID identification measure to thefinal user. This measure will be an effective way to successfully realize security of thenetwork and information system. Additionally, it uses the combination of fingerprintand IC card to collect and store the information. The ID identification of fingerprintand IC card is to deposit fingerprint's feature information and other privateinformation into an IC card as the basis to identification, ID identification mainlybases on the fingerprint identification of one to one or one to many, it could also useother information saved in the IC card as an assistant.Besides, this paper carries out fingerprint's identification in an embeddedsituation and hardware realization of the unsymmetrical encryption algorithms.Aiming at the characters of embedded algorithms, SOC chip from Arm series isselected as the core processor in this system, and this paper analyzed the structure andfunctions of this chip. The double CPU structures are used in the identificationtemplate, the major identification chip adopts SAMSUNG ARM9 S3C2500, theassistant chip adopts ATMEGA16L 8AI, which could collect images andcommunicate, The embedded system is mainly composed by several steps likefingerprint images' input, analysis and contrast, identification, communication withserial port and data saving, etc., and includes the selection and customization of theembedded system. And the software and hardware of the computer control part areconsisted of: the design of fingerprint's parallel port collected template and itsoriginal procedure, foreground fingerprint files' construction and contrastableprocedure, local area network's communicative procedure, data process procedure,and so on. This method also finishes a development course of online system tofingerprint identification, which introduces the basic theories of fingerprintidentification and the embedded TCP/IP contract's data transmission techniques innetwork communication. NS32490C, the selectable network card chip, makes theSCM existence in LAN and its data transmission in Intel network come true. The finaluser takes SCM as a media and makes fingerprint identification through network and telecommunications.According to national standard of Vehicle Traveling Data Recorder and concretepractical criterion of Shaanxi, this article introduces the application of the VehicleTraveling Data Recorder system, including the recorder host and data analysissoftware and the design of the embedded host. In this article, general situation of thesubject is presented firstly, then the system design and functions of recorder. On thebasis of these, it also describes, combined with the work experience, the process ofdesign and realization of data analysis software. Furthermore, this paper makes fartherdiscussion on the application of the biological identification techniques on the manageboard and on the vehicle assistant drive system. After testing, Vehicle Traveling DataRecorder works well and has a stable performance, which reaches the nationalstandard.