At MIT, Lorenzo Turicchia developed a distinctive patent portfolio spanning speech and audio processing, neural and sensory prostheses, wearable physiological monitoring, and low-power biomedical electronics. Together, these patent families demonstrate a coherent technical vision focused on biologically inspired, feedback-aware signal processing and its translation into practical low-power systems for hearing, speech, sensing, and medical monitoring.
In the current citation audit, the portfolio shows 133 citation occurrences from 129 unique later patent and publication numbers, associated with 38 companies and research organizations, plus 4 individually assigned citing patents.
These patent families have attracted later patent citations from major companies, medical-device firms, and research institutions, including Microsoft, Samsung, Boeing, Qualcomm, IBM, Dolby, Philips, Broadcom, Mitsubishi Electric, Masimo, ZOLL Medical, Fresenius Medical Care, Advanced Bionics, Second Sight, Brain Corporation, MED-EL, Cornell University, Princeton University, Georgia Tech Research Corporation, the University of Illinois, the University of Texas System, and the Bionics Institute of Australia, highlighting broad technical relevance across consumer electronics, aerospace, medical devices, physiological monitoring, hearing technologies, neural prostheses, and academic research.
Inventors: Lorenzo Turicchia, Rahul Sarpeshkar
Assignee: Massachusetts Institute of Technology (MIT)
Patent Family: US7787640B2; US20040252850A1; WO2004097799A1; EP1618559A1
Issue Date: August 31, 2010
Cited By (Companies / Organizations): Microsoft; Samsung; Philips; Dolby; Qualcomm; LG; Broadcom; Ericsson; Fraunhofer; SRS Labs; MStar Semiconductor; Advanced Bionics; MERL; Mitsubishi Electric; Sennheiser; Tencent; University of Illinois; MicronasNIT; etc.
Patent Influence: 21 citation occurrences across the patent family.
Technology Area: spectral enhancement; speech and audio processing; hearing-assistive signal processing; low-power auditory systems
Abstract: A spectral enhancement system is disclosed that includes an input node for receiving an input signal, at least one broad band pass filter coupled to the input node and having a first band pass range, at least one non-linear circuit coupled to the filter for non-linearly mapping a broad band pass filtered signal by a first non-linear factor n, at least one narrow band pass filter coupled to the non-linear circuit and having a second band pass range that is narrower than the first band pass range, and an output node coupled to the narrow band pass filter for providing an output signal that is spectrally enhanced.
Summary: This invention proposes a signal-processing architecture that enhances spectral components of signals, especially audio, through a combination of nonlinear mapping and multistage bandpass filtering. By applying broadband processing followed by narrower-band spectral shaping, the system enhances selected spectral components in a way that is relevant to speech clarity, sound quality, and intelligibility. The architecture is well aligned with auditory and hearing-related applications, including speech enhancement, hearing-assistive systems, and biologically inspired audio processing. The patent is also relevant to implementations where efficient or compact architectures are valuable, including low-power and miniaturized systems.
Unique Technical Features: broadband bandpass filtering; nonlinear compression and expansion; narrowband spectral enhancement following broader-band processing; multi-stage spectral shaping; biologically inspired auditory-processing concepts; suitability for compact and low-power implementation
Keywords: spectral enhancement, compression, expansion, cochlear modeling, hearing aid, low power, non-linear filter, adaptive audio processing, biomedical signal, auditory prosthesis
Notable Applications & Commercial Impact: This invention is highly relevant to later developments in speech enhancement, auditory signal processing, and hearing-related electronics. The later family-citation record shows continued downstream relevance in technologies associated with audio codec post-filtering, dialogue enhancement, multi-channel audio enhancement, speech intelligibility processing, and spectral contrast enhancement. Later citations associated with organizations such as Microsoft, Samsung, Philips, Dolby, Qualcomm, and LG further support the importance of the disclosed concepts in subsequent commercial and technical development.
Inventors: Lorenzo Turicchia, Soumyajit Mandal, Rahul Sarpeshkar
Assignee: Massachusetts Institute of Technology (MIT)
Patent Family: US14980109P; US12/700,214; US20100198094A1; US8708923B2
Issue Date: April 29, 2014
Cited By (Companies / Organizations): Microsoft; Qualcomm; Philips; Fresenius Medical Care; Masimo; Infosys; ZOLL; Hipass Design; University of Texas System; Georgia Tech Research Corporation; etc.
Patent Influence: 28 citation occurrences across the patent family.
Technology Area: wearable physiological monitoring; remote health monitoring; low-power biomedical sensing
Abstract: A wearable system for monitoring a plurality of physiological signals is provided. The wearable system includes at least one sensor producing the physiological signals associated with a patient. A processor unit receives the physiological signals from the at least one sensor. The processor unit analyzes the physiological signals to determine the occurrence of a triggered event and produces at least one output signal identifying the triggered event. A transmission unit receives the at least one output signal and prepares for transmission of the at least one output signal.
Summary: This patent introduces a compact, body-worn system capable of continuously measuring and analyzing multiple physiological signals—such as heart rate, motion, or respiratory activity. The system integrates sensors, a processor unit for real-time event detection, and a wireless transmission module for alerting healthcare providers. It is optimized for long-term, ambulatory monitoring without restricting patient mobility. The architecture emphasizes low power consumption and modular signal-processing to accommodate diverse sensor inputs. The technology anticipates personalized and preventive healthcare, particularly in outpatient and telemedicine scenarios.
Unique Technical Features: multi-sensor physiological integration; real-time event detection; wireless transmission module; low-power signal processor; wearable modular platform; continuous ambulatory monitoring.
Keywords: physiological monitoring, wearable sensor, remote health, ambulatory system, body-worn electronics, low power, healthcare telemetry, multi-signal analysis, patient tracking
Notable Applications & Commercial Impact: This invention contributed to the rise of wearable health monitoring systems capable of transmitting early alerts in real time. It helped anticipate a broader class of smart health patches, chest straps, and wrist-worn diagnostic or monitoring devices used in both consumer fitness and clinical settings. Its combination of multi-sensor acquisition, on-body event detection, and wireless transmission aligns with the technological direction later seen in connected health-monitoring systems, remote patient observation, and advanced medical telemetry platforms. The later citing records associated with companies and organizations such as Microsoft, Qualcomm, Philips, Fresenius Medical Care, Masimo, ZOLL, the University of Texas System, and Georgia Tech Research Corporation further indicate continued technical relevance of the disclosed concepts in adjacent commercial, academic, and medical-device development.
Inventors: Rahul Sarpeshkar, Lorenzo Turicchia, Soumyajit Mandal
Assignee: Massachusetts Institute of Technology (MIT)
Patent Family: WO2010017448A1; US20100036457A1; US8700166B2; EP2320830A1; EP2320830A4
Issue Date: April 15, 2014
Cited By (Companies / Organizations): Second Sight; Cornell University; The Bionics Institute of Australia; Brain Corporation; Qualcomm; SyncThink; GoPro; Princeton University; Bar-Ilan University; Pierre and Marie Curie University; Time Warner Cable; First Affiliated Hospital of Third Military Medical University; etc.
Patent Influence: 69 citation occurrences across the patent family.
Technology Area: visual prostheses; artificial vision; neural stimulation coding; sensory augmentation; low-power implantable biomedical systems
Abstract: A visual-prosthesis system codes visual signals into electrical stimulation patterns for the creation of artificial vision. In some examples, coding of the information uses image compression techniques, temporal coding strategies, continuous interleaved sampling (CIS), and/or radar or sonar data. Examples of the approach are not limited to processing visual signals but can also be used to process signals at other frequency ranges (e.g., infrared, radio frequency, and ultrasound), for instance, creating an augmented visual sensation.
Summary: This patent presents a method for transforming visual or non-visible signals into electrical stimulation patterns for artificial or augmented vision. The disclosed approach draws on image compression, temporal coding, continuous interleaved sampling, and radar/sonar-based sensing to form stimulation patterns that can be delivered through electrode arrays. The architecture is notable for extending beyond visible-light imaging to other frequency domains such as infrared, radio frequency, and ultrasound, broadening its relevance to sensory augmentation as well as prosthetic vision. The coding strategy is also well suited to implantable or power-constrained biomedical systems, where efficient representation and transmission of information are especially important.
Unique Technical Features: multispectral signal encoding; prosthetic vision coding; temporal-spatial compression; CIS-based stimulation presentation; radar/sonar-based sensory augmentation; cochlear-inspired coding concepts; neural stimulation interface; selective image-portion processing; transform-based coding with DCT coefficients.
Keywords: visual prosthesis, artificial vision, signal coding, neural interface, image compression, CIS, multispectral perception, retinal stimulation, biomedical implant
Notable Applications & Commercial Impact: This patent stands out for combining prosthetic-vision coding with broader sensory-augmentation concepts, allowing not only visual images but also infrared, radio-frequency, ultrasound, and radar/sonar-derived information to be converted into neural stimulation patterns. That makes the family relevant not only to retinal or thalamic visual prostheses, but also to next-generation assistive systems that translate non-visible environmental information into useful percepts. The later citation record connects the invention to organizations such as Second Sight, Cornell University, and The Bionics Institute of Australia, reinforcing its relevance to downstream work in retinal stimulation, image-contrast enhancement for prosthetic vision, and neural-coding strategies for artificial vision systems.
Inventors: Kenneth H. Wee, Lorenzo Turicchia, Rahul Sarpeshkar
Assignee: Massachusetts Institute of Technology (MIT)
Patent Family: WO2009023807A1; US20100217601A1; US8688438B2
Issue Date: April 1, 2014
Cited By (Companies / Organizations): ClearOne; etc.
Patent Influence: 1 citation occurrence across the patent family.
Technology Area: speech processing; speech synthesis; speech recognition; biologically inspired vocal-tract modeling; low-power voice processing
Abstract: A speech processing system includes a plurality of signal analyzers that extract salient signal attributes of an input voice signal. A difference module computes the differences in the salient signal attributes. One or more control modules control a plurality of speech generators using an output signal from the difference module in a speech-locked loop (SLL), the speech generators use the output signal to generate a voice signal.
Summary: This invention introduces a biologically inspired speech-processing architecture built around a speech-locked loop (SLL), in which signal analyzers extract salient attributes of an input voice signal, a difference module computes an error or mismatch, and control modules drive one or more speech generators in a feedback configuration. The patent is distinctive in linking speech analysis and speech generation through a closed-loop control architecture modeled on vocal production and auditory feedback. The disclosed approach is relevant not only to speech synthesis, but also to speech recognition, noise handling, feature extraction, and arbitrary sound processing, including operation beyond the audio band in some embodiments. The architecture is especially notable for implementations emphasizing low power, low hardware complexity, and analog or mixed-signal realizations.
Unique Technical Features: speech-locked loop (SLL); adaptive feedback synthesis; signal-attribute difference modules; biologically inspired speech-production modeling; analog vocal-tract implementation; cochlea-like speech analysis; multi-generator / multi-noise-generator feedback architecture; feedforward-plus-feedback control; applicability beyond the audio range.
Keywords: speech synthesis, adaptive speech loop, signal analyzer, real-time feedback, audio prosthetics, cochlear modeling, speech generator
Notable Applications & Commercial Impact: This patent stands out because it does not treat speech analysis and speech generation as separate blocks; instead, it unifies them in a biologically inspired feedback architecture that can lock generated speech to target speech attributes. That makes the family conceptually important for robust speech synthesis, speech imitation, articulatory modeling, adaptive voice generation, and noise-aware speech processing. The patent also extends beyond traditional synthesis by describing applications to speech recognition, training-sequence generation, arbitrary sound processing, and even higher-frequency operation outside the usual audio band. The later family-level citation by ClearOne connects the invention to downstream work in speech/noise processing and communications-oriented audio enhancement, while the patent text itself shows unusually broad conceptual reach for low-power, feedback-driven speech systems.
Inventors: Bharathwaj Ramakrishnan, Brian Schmidt-Nielsen, Lorenzo Turicchia, Rahul Sarpeshkar
Assignee: Mitsubishi Electric Research Laboratories, Inc. (MERL)
Patent Family: US20080027708A1; US7672842B2
Issue Date: March 2, 2010
Cited By (Companies / Organizations): IBM; Dolby; Ubercord; 41st Research Institute of CETC; etc.
Patent Influence: 6 citation occurrences across the patent family.
Technology Area: automatic speech recognition; speech enhancement; FFT-based signal processing; noise-robust speech front ends; low-power speech preprocessing
Abstract: A method and system processes a speech signal. A fast Fourier transform is performed on a speech signal to produce a speech signal having a plurality of frequency bands in a frequency domain. For each frequency band, the speech signal is filtered in the frequency domain with a spatial broadband filter, the broadband-filtered speech signal is compressed, the compressed speech signal is filtered with a spatial narrowband filter, and the narrowband-filtered signal is expanded.
Summary: This invention introduces an FFT-based companding framework for processing speech signals before automatic speech recognition. The disclosed method operates directly in the frequency domain, where each frequency band is processed through broadband filtering, compression, narrowband filtering, and expansion. The patent is especially notable because it adapts biologically inspired companding ideas to a computationally efficient FFT-based implementation, making the approach practical for real-time ASR front ends. By improving robustness to noise while remaining efficient, the architecture is relevant to embedded, low-power, and real-time speech-recognition systems.
Unique Technical Features: FFT-based spectral companding; frequency-domain broadband and narrowband filtering; band-specific compression and expansion; frame-based instantaneous processing; biologically inspired masking and tone-suppression behavior; direct extraction of recognition features from expanded signals; efficient real-time ASR front-end implementation.
Keywords: automatic speech recognition, FFT companding, speech enhancement, embedded voice interface, frequency filtering, low-power DSP
Notable Applications & Commercial Impact: This patent stands out because it translates biologically inspired auditory companding into an FFT-based architecture that is efficient enough for real-time speech-recognition front ends. That makes it especially relevant to noise-robust ASR, embedded voice interfaces, and low-power speech-processing pipelines where computational efficiency matters. The later citation record links the family to organizations such as IBM and Dolby, supporting continued downstream relevance in speech-feature extraction, coded-audio enhancement, and signal-processing methods related to robust audio analysis and recognition.
Inventors: Rahul Sarpeshkar, Lorenzo Turicchia
Assignee: Massachusetts Institute of Technology (MIT); Advanced Bionics LLC
Patent Family: US7415118B2; US20040136545A1; WO2004010417A2; PCT/US2003/022795; EP1529281B1; ATE347163T1; AU2003256653A1; CA2492246A1; DE60310084T2 (multi-jurisdiction family spanning the U.S., PCT/WO, Europe, Austria, Australia, Canada, and Germany)
Issue Date: August 19, 2008
Cited By (Companies / Organizations): Boeing; Samsung; Qualcomm; MED-EL; Rockstar Consortium; etc.
Patent Influence: 8 citation occurrences across the patent family.
Technology Area: spectral enhancement; distributed gain control; cochlear-inspired signal processing; hearing-assistive signal processing; low-power auditory front ends
Abstract: In accordance with an embodiment, the invention provides a spectral enhancement system that includes a plurality of distributed filters, a plurality of energy distribution units, and a weighted-averaging unit. At least one of the distributed filters receives a multi-frequency input signal. Each of the plurality of energy-detection units is coupled to an output of at least one filter and provides an energy-detection output signal. The weighted-averaging unit is coupled to each of the energy-detection units and provides a weighted-averaging signal to each of the filters responsive to the energy-detection output signals from each of the energy-detection units to implement distributed gain control. In an embodiment, the energy detection units are coupled to the outputs of the filters via a plurality of differentiator units, preferably double differentiation units.
Summary: This patent introduces a distributed gain-control architecture for spectral enhancement in multi-frequency signal-processing systems. Rather than applying gain control independently within isolated channels, the invention uses a spatially weighted interaction among neighboring filter outputs, allowing the collective system to adapt gain in a coordinated way. The disclosed architecture is especially notable for cochlear-inspired implementations, where distributed control helps reproduce biologically relevant frequency-response behavior while preserving wide dynamic range, temporal resolution, and low power consumption. The patent is particularly relevant to spectral enhancement in speech, auditory, and hearing-related systems, including cochlear-implant front ends and other low-power signal-processing platforms.
Unique Technical Features: distributed filter bank with shared gain control; energy-detection units coupled across neighboring channels; weighted spatial averaging kernel; nonlinear gain unit responsive to distributed energy estimates; optional differentiator units, including double differentiation; cochlear-inspired filter-cascade implementation; masking and two-tone suppression behavior; tradeoff between spectral contrast preservation and audibility; low-power front-end suitability.
Keywords: distributed gain control, spectral enhancement, cochlear model, silicon cochlea, hearing aid, cochlear implant, low power, masking, auditory signal processing, nonlinear filtering
Notable Applications & Commercial Impact: This patent stands out because it frames gain control as a distributed, system-level phenomenon rather than a purely local or channel-by-channel adjustment. That makes the family especially relevant to cochlear-inspired signal processing, spectral enhancement in noisy environments, and hearing-assistive front ends where preserving both audibility and spectral contrast is critical. The patent text explicitly positions the architecture as advantageous for cochlear-implant processors and for implementing computationally intensive biological-cochlea algorithms with very low power. The later citation record, including citations associated with Boeing, Samsung, Qualcomm, and MED-EL, further supports downstream relevance in speech enhancement, intelligibility improvement, audio equalization, amplifier control, and channel-specific gain-control technologies for auditory devices.
Inventors: Lorenzo Turicchia, Rahul Sarpeshkar
Assignee: Massachusetts Institute of Technology (MIT)
Patent Family: U.S. provisional patent application
Technology Area: cardiovascular signal processing; physiological monitoring; feedback-based biomedical signal analysis; low-power medical electronics
Summary: This invention centers on the use of feedback mechanisms in cardiovascular signal-processing systems, combining biomimetic signal interpretation with adaptive feedback control to improve the accuracy, responsiveness, and robustness of cardiovascular monitoring. The concept is well aligned with heart-rate, ECG, blood-pressure, and related physiological-signal applications, where real-time feedback can improve signal interpretation under dynamic conditions. The work also fits naturally within a broader framework of wearable and implantable physiological-monitoring systems, with particular relevance to ultra-low-power implementation, continuous monitoring, and adaptive biomedical electronics.
Unique Technical Features: biomedical feedback circuits; low-power analog processing; cardiovascular signal detection; adaptive control; real-time analysis
Keywords: cardiovascular monitoring, biofeedback, analog signal processing, wearable medical device, physiological signal loop
Notable Applications & Commercial Impact: This work is relevant to the development of wearable medical electronics and intelligent health-monitoring systems that combine physiological signal acquisition with adaptive filtering, interpretation, and response. Its emphasis on feedback-based cardiovascular signal processing suggests strong applicability to next-generation low-power medical devices designed for continuous physiological monitoring, robust signal tracking, and improved responsiveness in real-world biomedical settings.