MORGAN STATE UNIVERSITY

An Automated Computer-Vision based system and method for identifying potential hazards to building occupants in building exit pathways is described to address the inherent challenges in manual building safety inspections. The computer vision-based approach automates the inspection of an interior building hallway (exit access) for an obstruction that may be a potential hazard. The invention mitigates the risk of a hazard to the building occupants by sensing and alerting the safety officer before a situation turns into an emergency. Test results demonstrate that the Dual Temporal Buffer Differencing (DTBD) method of the invention can detect a potential hazard in the building exit access effectively and continuously. As a result, the invention can facilitate safety in the building and allow safety inspectors to make better use of resources.

An integrated and autonomous mobility system for disabled persons including an automated wheelchair (AV) and an Adapted Automated Vehicle (AAV) configured to work together as an integrated unit to transport a user to a pre-selected destination with little to no input from the user during the trip. The AV is also configured to operate independently of the AAV to transport the user on shorter trips, or to continue a journey after the AAV has transported the AW to an intermediate destination. Multiple safety systems are included. The system has an override function according to which a user can interrupt the automated transport operation and cause the AV and/or AAV to stop, slow down, speed up, or change direction.

This invention relates to the development of heat-tolerant soft-shell clam (Mya arenaria) lines for aquaculture through selective breeding and molecular genetic approaches. Two methods are employed: (1) phenotype-based selection, in which clams are exposed to natural or controlled heat shock conditions (30-33° C.) and survivors are selected across successive generations and multiple populations; and (2) marker-assisted selection (MAS), utilizing single nucleotide polymorphisms (SNPs) identified through genome-wide association studies (GWAS) to select broodstock capable of producing heat-tolerant offspring. The resulting lines exhibit enhanced survival and endurance under elevated temperatures, particularly above the species' critical threshold of 28° C. These heat-tolerant clams enable sustainable aquaculture production in Maryland and other warm, low-salinity estuaries where soft-shell clams historically failed to survive summer heat.

This invention relates to genetically improved Eastern oyster ( Crassostrea virginica ) lines developed by the Morgan State University Patuxent Environmental and Aquatic Research Laboratory. Derived from Maryland wild populations, these diploid, triploid, and tetraploid lines exhibit enhanced tolerance to low salinity, rapid growth, and disease resistance. Genetic improvement was achieved through multiple approaches including phenotype-based, genomic, and marker-assisted selection. Several diploid low-salinity-tolerant lines (LS2019, LS2025, LS H-GEBV) and a fast-growing line (FG H-GEBV) have been produced and are undergoing field evaluation, while MSX-resistant lines are under development through field challenge testing. Multi-trait genomic selection enables integration of growth, salinity tolerance, and disease resistance into single lines based on trait correlations and production needs. Triploid lines (3nD-20, 3nE-20, 3 nF-20, 3nW-25) generated by chemical induction are used to create tetraploids, ultimately producing superior commercial triploid seed. These distinct oyster lines improve aquaculture productivity and restoration success in low-salinity Chesapeake Bay environments.

A compact portable quantum sensor device consists of a very high efficiency RF antenna, a low power RF signal generator, an RF amplifier (if needed), an RF switch and isolator, a photodetector(s) with amplifier circuit, various optical components, a microcontroller, a Bluetooth communication board, a battery, a display, and control software and machine learning algorithms for noise reduction and data analysis. A battery-operated portable quantum sensor device prototype was built for magnetometry applications. A second prototype with a smaller sensor head footprint was also built for brain imaging through measuring magnetic fields induced by brain neuron action potentials. In a third photonic circuit approach, all of the optical and electronic components are integrated on the same chip.

A method for preparing robust pharmaceutical shaped bodies, made thereby for delayed drug release, wherein these shaped bodies are also stable in physiological media from few days to several weeks. The method includes adding powders containing active ingredients and support material components, wherein the support material is preferably metal-organic framework, adding the additional liquid components (binders), homogenization of the obtained mixture with the use of high-speed mechanical mixing, pressing to the pharmaceutical shaped bodies, and drying these shaped bodies. The obtained pharmaceutical shaped bodies can be used for long-term (from few days to several weeks) delayed release of medicinal drugs.

Methods and systems identify a device on an IoT network. A network controller detects connection of the device on the IoT network and determines an organizational unique identifier (OUI) from a media access control (MAC) address of the device. The network controller retrieves a vendor ID from an OUI database. The network controller sends requests to the device via transmission control protocol (TCP) ports and receives a TCP port output. The network controller sends user datagram protocol (UDP) datagrams to the device via UDP ports and receives a UDP port output. The network controller determines a device fingerprint from the vendor ID, the TCP port output, and the UDP port output and matches the device fingerprint to parameters in an identity database to determine a first identity of the device. The network controller sends a DHCP deauthorization request to the device. After detecting reconnection of the device on the IoT network, the network controller collects DHCP option 55 parameters from the device and refines the first identity to determine a second identity based on the parameters.

Methods provide device authentication for an intrusion detection system implementing building automation and control network (BACnet) Master-Slave/Token-Passing (MS/TP). An authentication protocol provides countermeasures to vulnerabilities in the BACnet MS/TP physical layer by utilizing an extended message format to cloak device identifiers (IDs). Adversaries are prevented from using known device IDs to gain access to the network. An authenticating device hashes a device identifier of a device to be authenticated combined with a random number. The authenticating device receives a hash of the random number plus the device identifier from the device. The authenticating device compares the hashes and authenticates the device if the hashes match. To transmit the hash, the BACnet MS/TP frame format includes an extended header cyclic redundancy check (CRC) field having bytes reallocated from the data field of the frame format. Another countermeasure utilizes a physical unclonable function (PUF) of the device in the extended header CRC.

A device and method for early detection of eye disease including a corneal topography device configured to measure the elevation of a patient's anterior and posterior corneas sequentially or simultaneously; organize the elevation data for each cornea into a two-dimensional matrix where the center of the cornea is in the center of the data frame, rotate the data for a first eye 180 degrees around the Y axis relative to a second eye, subtract data on each corresponding corneal point, and generate an elevation difference matrix showing the degree of symmetry or asymmetry between the patient's left and right eye corneal topography. A quantitative corneal symmetry index that measures the three-dimensional (3D) volume difference between the corneal surfaces of fellow eyes for differentiating between normal and abnormal corneas, including early detection and prognosis of ectatic disorders such as keratoconus and Fuchs dystrophy.

A data anomaly detection method and apparatus in which a deep neural network is trained on baseline data. Sequences of statistics of each layer of the deep neural network are saved, processed and used to train an LSTM autoencoder across a variety of reconstruction error thresholds, and a preferred threshold is selected for an optimized autoencoder. In an Inference mode, a data sample is presented to the autoencoder; the reconstruction error is calculated and compared to the threshold. If it is above the threshold, then the data sample is an out-of-distribution sample, and the sample is tagged as anomalous.