Vincent De Sapio

An approach for goal-oriented neuromuscular control of digital humans in physics-based simulations

We present a novel approach for the control of digital humans modelled as musculoskeletal systems in physics-based environments. This approach uses high-level goal-oriented commands as input, from which low-level neuromuscular excitations are produced that generate motion consistent with the high-level command input. Central to our approach is the reformulation of a neuromuscular control algorithm in task space, a space defined by coordinates relevant to the specific task being performed, rather than the full configuration space of the digital human. A control methodology is also detailed for addressing holonomic constraints in the skeletal kinematics. Examples are presented that demonstrate the motion control architecture for an arm-reaching simulation.

Neuromorphic and early warning behavior-based authentication for mobile devices

Finding the balance between security, privacy, and usability for mobile authentication has been an active area of research for the past several years. Many researchers have taken advantage of the availability of multiple sensors on mobile devices and have used these data to train classifiers to authenticate users. For example, implicit authentication algorithms have been developed based on behavior patterns identified from a combination of sensors including location, co-location, application usage, biometric measurements, continuity of interaction between the user and the phone, and possession of the phone [1,2,3,4,5]. Furthermore, the onboard sensors of mobile devices have previously been used to identify users based on touch [6] and fusions of touch and speech inputs [7]. However, a system utilizing low-power onboard electronics for anomaly detection and user classification is lacking. Here, we report on the performance of two subsystems tested in a controlled use scenario to classify and authenticate users of a mobile device. The overall system utilizes two subsystems for anomaly detection and user classification: (1) a neuromorphic chip for continuous, low-power, online monitoring and classification, and (2) an early warning system (EWS) algorithm for longer duration time-series behavioral and biometric classification.

Human Factors Simulation Using Demographically Tuned Biomechanical Models

We have developed a set of tools based on the OpenSim simulation framework that allows for input of anthropometry, muscle geometry, and measured strength capability to generate demographically tuned models from a core generic musculoskeletal model. The strength tuning capability exploits an algorithm that adaptively tunes muscle parameters in generic Hill-type muscle models to generate performance data consistent with ergonomic subject studies of specific demographic populations (e.g. elderly populations). Following the tuning of the generic model to generate a demographic-specific model, human performance in a variety of scenarios can then be analyzed. Currently, the model is in a prototype phase and has been applied to scenarios modeling elderly passengers interacting with airplane interior features including overhead bins and lavatories. The next phase of development will include manufacturing scenarios with input based on motion capture and worker demographics, including strength measurements.

Neuromorphic and Early Warning behavior-based authentication in common theft scenarios

With increased rates of smartphone theft over the past decade, mobile authentication systems that operate on a continual basis are a necessity to meet increasing demands for user privacy, device usage, and authentication accuracy. Rather than forcing end users to continually self-authenticate via password pins or through other means on a time-interval basis [2–7], a system that continuously authenticates users provides a more frictionless relationship between a users device and its physical security. Such a system, if effectively operating on a low-powered, unobtrusive, and secure basis, would make it viable for most consumer mobile devices. In this paper, we build upon our work in [1] to provide a novel authentication scheme that meets these requirements for a commonly adopted system. Our system, iSentinel, hopes to provide an unobtrusive, low-powered solution for detecting and responding to common theft scenarios by continuously authenticating mobile devices in use cases such as walking, texting, and driving.

An approach and implementation for coupling neurocognitive and neuromechanical models

The neuromechanics of human motion are generally represented in the literature by feedforward control mechanisms: the brain sends a control signal to a part of the body to move, and motion ensues. Thus neuromechanical commands for motion are influenced by control signals from neurocognitive inputs. However, feedback also exists from the neuromechanical system to the neurocognitive one, so that subsequent decisions related to motor commands are influenced by the motion itself. Recent work suggests that accounting for bidirectional feedback, both from neurocognitive to neuromechanical systems and from neuromechanical to neurocognitive ones, allows for more robust accounts of behavior. In this paper, we describe a neurocognitive model, a neuromechanical model, and a simple bidirectional feedback mechanism to couple the two systems. We then show that the behavior of the combined system is determined by the interaction of coupling strength and properties of the mechanical and cognitive models.