A mathematical formalization of the problem of acoustic interference in swarm monitoring systems is presented. Existing approaches are described that solve the problem of suppressing the own noise of a single drone, but do not take into account dynamic acoustic channels and the unsteadiness of the spectrum during maneuvers. A signal model is proposed that takes into account the above factors and serves as a theoretical basis for the development of cooperative noise reduction algorithms. An analysis of the model practical applicability for swarms of several drones is carried out.

For embodied agents to navigate and reason indoor spaces, they need object-level 3D representations that stay consistent over time as new frames arrive from a monocular camera. Current online 3D instance segmentation methods either depend on posed RGB-D input with ground-truth depth or couple tightly to the internal representations of specific foundation models, sacrificing modularity. We observe that appearance-based and geometry-based object matching exhibit complementary failure modes: appearance is ambiguous among spatially separated duplicates, while geometry is unreliable for visually distinct objects at similar locations. This motivates SAM3R, a training-free pipeline that fuses spatial overlap, 3D centroid displacement, and visual-semantic similarity into a single assignment cost solved via bipartite matching. The cost is constructed entirely from the outputs of frozen foundation models without accessing internal representations. Object tracks are classified through a cascaded decision tree that detects scene changes via field-of-view gated temporal voting. On ScanNet200 and Replica, SAM3R performs competitively with methods that require architecture-specific features or additional training, despite operating in a fully online, monocular setting. Qualitative evaluation on the Aria Digital Twin dataset further demonstrates that the pipeline maintains correct object identities through physical object manipulation, including hand occlusion and large spatial displacement.

The choice of controlled outputs divided into two subsets when organizing integrated control circuits for discrete devices using the properties of parity codes, are considered. Three basic computing control structures are described that operate in pulse mode using time redundancy and representing signals as sequences of zeros and ones. A structure for monitoring calculations at the outputs of discrete devices is proposed, combining the properties of the classical parity control structure and a structure with control of self-duality of calculations. The features of the distinguished subsets of outputs of discrete devices are established. The possibility is shown to identify methods of dividing outputs into subsets, for the control of each of which only self-dual or only self-quasi-dual Boolean functions “close” to them are used. Formulas describing the number of ways of dividing the outputs of discrete devices into subsets with special properties are obtained. Examples of calculations based on these formulas are given. The regularities inherent in the structures under consideration and manifesting themselves with an increase in the number of outputs of discrete devices are described. The results of the presented study may be of interest to developers of discrete devices with fault detection implemented on various element bases.

The problem of designing hardware accelerators for edge artificial intelligence systems is considered, taking into account the limitations on the resources used. A modular parameterizable architecture of the neural network processor is been developed, which allows flexibly adjusting the characteristics of the accelerator, including the size of computing elements and the size of the internal buffer memory, to the requirements of the target system. The effectiveness of the obtained results is confirmed by modeling a neural network accelerator at the level of register transfers using the parameters of the convolutional layer of the YOLOv5s neural network. The dependence of the crystal area on the architecture parameters using the programmable logic unit (Field Programmable Gate Array, FPGA) of the ZYNQ-7000 system on a chip is investigated. It has been shown that increasing the size of the systolic array increases productivity, but the effect of the increase decreases with large values. Scaling the systolic array also increases the number of logic gates and single-bit registers of the FPGA. Increasing the amount of internal memory significantly affects the consumption of FPGA block memory elements, but does not significantly affect overall performance. The experimental results confirmed the advantage of the proposed approach in terms of adaptability and flexibility of customization for various application scenarios. Promising areas of application are systems that require high performance with limited resources. The results obtained can serve as a basis for the development of new generations of neural network processors with the ability to scale performance and taking into account resource constraints.

An algorithm of computational design for finding optimal parameters for rotary drives of locomotion robots that meet performance and behavioral requirements, is presented. Anthropomorphic robots, quadrupeds, and so-called general-purpose robots are equipped with pedipulatory mechanisms for navigating unstructured and a priori unknown environments. The requirements imposed on the mechanisms in terms of load capacity, dynamics, energy efficiency, safety and aesthetics may contradict each other. For example, aesthetics involves minimizing the overall dimensions of the drives, the load capacity requires an increase in gear ratio, and this increases reflected inertia and friction losses, adversely affecting dynamics and energy efficiency. In the external cycle of the proposed algorithm, the optimal parameters of rotary drives are searched for by the total weight and gear ratio of the drives. Regression models trained on the passport data of commercial drives recover key parameters such as nominal and peak torques, reflected inertia, motor constant, estimated overall dimensions of the drives and the efficiency of the gearbox. In the internal cycle of the proposed algorithm, the trajectory optimization problem is solved for a complete model of robot dynamics with limitations on torque, velocity, kinematics and contacts by estimating energy through an electrical model with separation of electrical losses and friction losses. The proposed algorithm is tested to solve the problem of designing a quadroped. The algorithm’s input data include behavioral characteristics (walking, dynamic movements such as jumping) and performance parameters (for example, when handling a payload). The algorithm’s output is a set of physical parameter estimates sufficient for design considerations. Verification of the parameters is performed using simulation tools.

With the help of robotic anthropomorphic hands capable of performing functional grips, it is possible to use hand tools and everyday objects for their intended purpose, which eliminates the need for specialized equipment. The choice of design parameters and working surfaces of tactile sensors is justified. A method of system design of such hand with tactile feedback is presented. First, a discrete space of design candidates is formed based on the characteristics of the drives and components. Then, planning and verification of trajectories for functional captures are performed; candidates are ranked according to a heuristic function reflecting capture efficiency metrics. Based on the results of modeling the best option, the dimensions of the working surfaces of the phalanges and the range of contact forces are determined; taxels are synthesized — elastic elements of the working surface of the tactile sensor. The parameters providing a compromise between the sensitivity of the taxiways and their number are investigated, and the relationship of the taxiway deformations with contact forces is estimated for each variant. Based on the results of a posteriori analysis, taking into account the requirements for the range of measured forces and resistance to incomplete data, the geometric parameters of the sensor are selected. The method is tested by designing and testing a prototype of an anthropomorphic robotic hand.

The choice of controlled outputs divided into two subsets when organizing integrated control circuits for discrete devices using the properties of parity codes, are considered. Three basic computing control structures are described that operate in pulse mode using time redundancy and representing signals as sequences of zeros and ones. A structure for monitoring calculations at the outputs of discrete devices is proposed, combining the properties of the classical parity control structure and a structure with control of self-duality of calculations. The features of the distinguished subsets of outputs of discrete devices are established. The possibility is shown to identify methods of dividing outputs into subsets, for the control of each of which only self-dual or only self-quasi-dual Boolean functions “close” to them are used. Formulas describing the number of ways of dividing the outputs of discrete devices into subsets with special properties are obtained. Examples of calculations based on these formulas are given. The regularities inherent in the structures under consideration and manifesting themselves with an increase in the number of outputs of discrete devices are described. The results of the presented study may be of interest to developers of discrete devices with fault detection implemented on various element bases.

A combined model of synchronization of digital and physical processes for the production of small CubeSattype spacecraft taking into account the degree of automation and the specifics of each stage of production in order to achieve the least possible desynchronization, is developed. The methodological basis is the methods of synchronization of digital and physical production processes, which ensure that the dynamics of controlled operations are taken into account and predicted. Using Python, the misalignment is calculated using several synchronization models for each stage of production, as well as for the total average error of the entire production. The mathematical model of combined synchronization is adjusted by introducing weighting factors that take into account the degree of automation of the CubeSat manufacturing process, considering the advantages of each individual model. The proposed model can be used in processes of various levels of automation: from low to close to the serial production level of small spacecrafts.