Typical Military Applications of Speech Recognition Technology

Author: Military Eagle Think Tank Source: Military Eagle Dynamics

Language and military operations have always had a natural connection. Generally, military activities such as organization and command, propaganda, enemy communication, and code interpretation cannot be separated from human language, especially natural human speech. The ability of speech recognition has always been an important component of military intelligence automatic processing capabilities. The high-tech development strategies of military powers, such as the US Department of Defense’s AVLEY program, the US Defense Advanced Research Projects Agency’s GALE project, and the EU’s ESPRIT program, have all included it in their research plans, investing significant human, material, and financial resources to enhance research on speech recognition and understanding, aiming to overcome the high-tech challenge of human-computer interactive natural speech communication. Semantic recognition technology not only provides a convenient human-computer interaction interface for remote military communication and remote control of battlefield robots but more importantly, as a typical representative of biometric technology, it utilizes human physiological characteristics as a confirmation of identity “mark”, which has a high security advantage unmatched by traditional identity verification methods, maximizing the security of national defense information and facilities, and can be widely applied in military confidentiality and command confirmation fields.

Due to technological limitations, traditional military command and control systems often require keyboard input to complete tasks such as threat assessment, situational evaluation, and assisting command decisions. However, manual keyboard input is slow, prone to errors, and can distract commanders’ attention, becoming a limiting factor in shortening the response time of military command systems. In contemporary high-tech communication systems, speech recognition is gradually becoming a key technology for human-computer interfaces. The combination of speech recognition technology with speech synthesis and understanding technologies enables users to operate directly through voice commands, eliminating the need for keyboards. For example, users can conveniently query and extract relevant information from remote database systems through voice commands, or remotely control battlefield robots or industrial robots. In many current applications, voice input has become an ideal means of information input, thereby making intelligent voice interfaces personify and intelligentize the “tool” attributes of communication equipment, transforming them into “providers” of services, with the simple and quick interaction methods they provide becoming a new interface for human-computer interaction in military communication.

Fully applying speech signal digital processing technology in military management, command, control, and communication, converting information input from keyboards to voice input, and minimizing the time wasted on manual operations is of great significance for further improving the automation level of military command systems and enhancing the rapid response capabilities of military commands.

Identity recognition and authentication are fundamental technologies for ensuring information security in the military field. In recent years, with changes in the environment and technological development, the demand for information security, financial security, and public safety has increased significantly. Traditional password-based identity recognition technologies (such as passwords and PIN codes) have become increasingly inadequate to meet the real needs of high-security industries such as national defense and finance.

The voiceprint recognition technology in speech recognition has important application value in military confidentiality. In the closed management of military computer systems and critical areas, using voiceprint recognition technology for identity authentication has a high degree of accuracy, which can further enhance system security. For example, some new computer security products that apply voiceprint recognition technology can add voiceprint authentication functions based on ordinary USB encryption keys and encrypt computer systems to protect important files from illegal theft, browsing, tampering, deletion, or destruction. This technology complies with national security standards, and the operations for file encryption and decryption are extremely simple, providing multiple layers of security protection, effectively preventing unauthorized users from accessing, using, and stealing computer systems. Additionally, in some critical areas of military installations, using speech recognition technology for access control management allows the confidentiality management system to authenticate identity based on the input natural speech signal and automatically open or close access facilities, effectively identifying legitimate entrants and exits.

In military operations, issuing commands via telephone is a common method of information transmission. By applying voiceprint recognition technology, it is possible to confirm the identity of the command issuer at the biometric level, avoiding situations where the enemy disguises as our commander to issue false commands and disrupt our military operations. As early as 2002, Microsoft had researched that traditional identity recognition technologies would gradually be replaced by biometric technologies and partnered with the US RSA company to establish a “Tamper-Resistant Biometric ID Card” program worth over $7 billion. Due to the conversion of recorded audio from analog signals to digital signals during computer information processing, and the playback process also undergoing a conversion from digital signals back to analog signals, even if eavesdroppers use recording devices to capture the legitimate user’s voice for voiceprint authentication, the voice spectrum will show significant attenuation and distortion after two conversions, making it easy for authentication programs to distinguish this distortion. Therefore, relying on recorded audio for login cannot pass voiceprint authentication. By appropriately adjusting the strictness threshold for voiceprint authentication, the “false acceptance rate” and “false rejection rate” of voiceprint authentication can be decreased.

The correct decisions of military commanders stem from a wealth of accurate intelligence. Under the current conditions supported by advanced technology, the abundance of channels and methods for obtaining intelligence has made distinguishing the authenticity of information a focus for military intelligence personnel. With the vast amount of audio-visual intelligence emerging, biometric recognition technologies, including voiceprint recognition, have begun to shine in the field of intelligence processing. Through public and eavesdropping channels, the United States has established a powerful data system that stores information on almost all world leaders. For example, the US intelligence agencies have set up nearly ten projects directly related to internet surveillance, and Germany’s Der Spiegel reported in 2014, citing documents provided by Snowden, that the NSA monitored 122 foreign leaders in 2009, in addition to collecting nearly 5 billion mobile call location records and about 2 billion global SMS messages daily. For the voice data collected from leaders of various countries, the CIA records and analyzes the voices for ease of use, converting them into spectrograms or voiceprints. The electronic filters in the spectrogram only allow specific frequency bands to pass through and record the intensity of these frequencies. For instance, during the Iraq War, whenever Al Jazeera broadcasted recordings of Saddam’s speeches, CIA voice recognition experts and language specialists would begin analyzing those recordings, comparing the voices on the tapes with stored voices to identify similarities and differences, which involve breathing patterns, tone variations, unusual speech habits, and regional dialects, thereby determining the authenticity of the recorded information. Military eavesdropping and identification are beneficial for quickly assessing the true value of acquired intelligence, and in wartime, they help commanders grasp battlefield information in a timely manner to make correct decision judgments.

Military intelligent robots are widely used in military fields, possessing certain sensory functions, capable of making independent decisions and taking corresponding actions based on the commands of the operators. Speech recognition technology and speech synthesis technology are important research directions in the field of military intelligent robots, having become key technologies for autonomous robot automatic control with significant application value in large-scale manufacturing control, battlefield environmental attacks, and rescue robot deployments.

Remote voice control of robots is based on speech recognition technology, aiming to enable robots to understand human language and carry out related actions based on heard commands, achieving various tasks for humans and facilitating communication between humans and machines. Currently, the remote voice control technology for industrial robots is relatively mature; however, there is still significant room for improvement in remote battlefield control under strong interference environments. The combination of speech recognition technology and robot control technology also reflects the automation of technology; speech recognition technology, as a vital link between human language and robot language, should receive more attention.