Brain-inspired multimodal hybrid neural network for robot place recognition

Reliable object tracking by multimodal hybrid feature extraction and transformer-based fusion

Visual object tracking, which is primarily based on visible light image sequences, encounters numerous challenges in complicated scenarios, such as low light conditions, high dynamic ranges, and background clutter. To address these challenges, incorporating the advantages of multiple visual modalities is a promising solution for achieving reliable object tracking. However, the existing approaches usually integrate multimodal inputs through adaptive local feature interactions, which cannot leverage the full potential of visual cues, thus resulting in insufficient feature modeling. In this study, we propose a novel multimodal hybrid tracker (MMHT) that utilizes frame-event-based data for reliable single object tracking. The MMHT model employs a hybrid backbone consisting of an artificial neural network (ANN) and a spiking neural network (SNN) to extract dominant features from different visual modalities and then uses a unified encoder to align the features across different domains. Moreover, we propose an enhanced transformer-based module to fuse multimodal features using attention mechanisms. With these methods, the MMHT model can effectively construct a multiscale and multidimensional visual feature space and achieve discriminative feature modeling. Extensive experiments demonstrate that the MMHT model exhibits competitive performance in comparison with that of other state-of-the-art methods. Overall, our results highlight the effectiveness of the MMHT model in terms of addressing the challenges faced in visual object tracking tasks.

Adaptive spatiotemporal neural networks through complementary hybridization

Processing spatiotemporal data sources with both high spatial dimension and rich temporal information is a ubiquitous need in machine intelligence. Recurrent neural networks in the machine learning domain and bio-inspired spiking neural networks in the neuromorphic computing domain are two promising candidate models for dealing with spatiotemporal data via extrinsic dynamics and intrinsic dynamics, respectively. Nevertheless, these networks have disparate modeling paradigms, which leads to different performance results, making it hard for them to cover diverse data sources and performance requirements in practice. Constructing a unified modeling framework that can effectively and adaptively process variable spatiotemporal data in different situations remains quite challenging. In this work, we propose hybrid spatiotemporal neural networks created by combining the recurrent neural networks and spiking neural networks under a unified surrogate gradient learning framework and a Hessian-aware neuron selection method. By flexibly tuning the ratio between two types of neurons, the hybrid model demonstrates better adaptive ability in balancing different performance metrics, including accuracy, robustness, and efficiency on several typical benchmarks, and generally outperforms conventional single-paradigm recurrent neural networks and spiking neural networks. Furthermore, we evidence the great potential of the proposed network with a robotic task in varying environments. With our proof of concept, the proposed hybrid model provides a generic modeling route to process spatiotemporal data sources in the open world.

Crossmodal sensory neurons based on high-performance flexible memristors for human-machine in-sensor computing system

Constructing crossmodal in-sensor processing system based on high-performance flexible devices is of great significance for the development of wearable human-machine interfaces. A bio-inspired crossmodal in-sensor computing system can perform real-time energy-efficient processing of multimodal signals, alleviating data conversion and transmission between different modules in conventional chips. Here, we report a bio-inspired crossmodal spiking sensory neuron (CSSN) based on a flexible VO2 memristor, and demonstrate a crossmodal in-sensor encoding and computing system for wearable human-machine interfaces. We demonstrate excellent performance in the VO2 memristor including endurance (>1012), uniformity (0.72% for cycle-to-cycle variations and 3.73% for device-to-device variations), speed (<30 ns), and flexibility (bendable to a curvature radius of 1 mm). A flexible hardware processing system is implemented based on the CSSN, which can directly perceive and encode pressure and temperature bimodal information into spikes, and then enables the real-time haptic-feedback for human-machine interaction. We successfully construct a crossmodal in-sensor spiking reservoir computing system via the CSSNs, which can achieve dynamic objects identification with a high accuracy of 98.1% and real-time signal feedback. This work provides a feasible approach for constructing flexible bio-inspired crossmodal in-sensor computing systems for wearable human-machine interfaces.

Nanowatt all-optical 3D perception for mobile robotics

Three-dimensional (3D) perception is vital to drive mobile robotics' progress toward intelligence. However, state-of-the-art 3D perception solutions require complicated postprocessing or point-by-point scanning, suffering computational burden, latency of tens of milliseconds, and additional power consumption. Here, we propose a parallel all-optical computational chipset 3D perception architecture (Aop3D) with nanowatt power and light speed. The 3D perception is executed during the light propagation over the passive chipset, and the captured light intensity distribution provides a direct reflection of the depth map, eliminating the need for extensive postprocessing. The prototype system of Aop3D is tested in various scenarios and deployed to a mobile robot, demonstrating unprecedented performance in distance detection and obstacle avoidance. Moreover, Aop3D works at a frame rate of 600 hertz and a power consumption of 33.3 nanowatts per meta-pixel experimentally. Our work is promising toward next-generation direct 3D perception techniques with light speed and high energy efficiency.

Memory-electroluminescence for multiple action-potentials combination in bio-inspired afferent nerves

The development of optoelectronics mimicking the functions of the biological nervous system is important to artificial intelligence. This work demonstrates an optoelectronic, artificial, afferent-nerve strategy based on memory-electroluminescence spikes, which can realize multiple action-potentials combination through a single optical channel. The memory-electroluminescence spikes have diverse morphologies due to their history-dependent characteristics and can be used to encode distributed sensor signals. As the key to successful functioning of the optoelectronic, artificial afferent nerve, a driving mode for light-emitting diodes, namely, the non-carrier injection mode, is proposed, allowing it to drive nanoscale light-emitting diodes to generate a memory-electroluminescence spikes that has multiple sub-peaks. Moreover, multiplexing of the spikes can be obtained by using optical signals with different wavelengths, allowing for a large signal bandwidth, and the multiple action-potentials transmission process in afferent nerves can be demonstrated. Finally, sensor-position recognition with the bio-inspired afferent nerve is developed and shown to have a high recognition accuracy of 98.88%. This work demonstrates a strategy for mimicking biological afferent nerves and offers insights into the construction of artificial perception systems.

Artificial Intelligence-Powered Electronic Skin

Skin-interfaced electronics is gradually changing medical practices by enabling continuous and noninvasive tracking of physiological and biochemical information. With the rise of big data and digital medicine, next-generation electronic skin (e-skin) will be able to use artificial intelligence (AI) to optimize its design as well as uncover user-personalized health profiles. Recent multimodal e-skin platforms have already employed machine learning (ML) algorithms for autonomous data analytics. Unfortunately, there is a lack of appropriate AI protocols and guidelines for e-skin devices, resulting in overly complex models and non-reproducible conclusions for simple applications. This review aims to present AI technologies in e-skin hardware and assess their potential for new inspired integrated platform solutions. We outline recent breakthroughs in AI strategies and their applications in engineering e-skins as well as understanding health information collected by e-skins, highlighting the transformative deployment of AI in robotics, prosthetics, virtual reality, and personalized healthcare. We also discuss the challenges and prospects of AI-powered e-skins as well as predictions for the future trajectory of smart e-skins.

Perceiving like a Bat: Hierarchical 3D Geometric-Semantic Scene Understanding Inspired by a Biomimetic Mechanism

Geometric-semantic scene understanding is a spatial intelligence capability that is essential for robots to perceive and navigate the world. However, understanding a natural scene remains challenging for robots because of restricted sensors and time-varying situations. In contrast, humans and animals are able to form a complex neuromorphic concept of the scene they move in. This neuromorphic concept captures geometric and semantic aspects of the scenario and reconstructs the scene at multiple levels of abstraction. This article seeks to reduce the gap between robot and animal perception by proposing an ingenious scene-understanding approach that seamlessly captures geometric and semantic aspects in an unexplored environment. We proposed two types of biologically inspired environment perception methods, i.e., a set of elaborate biomimetic sensors and a brain-inspired parsing algorithm related to scene understanding, that enable robots to perceive their surroundings like bats. Our evaluations show that the proposed scene-understanding system achieves competitive performance in image semantic segmentation and volumetric-semantic scene reconstruction. Moreover, to verify the practicability of our proposed scene-understanding method, we also conducted real-world geometric-semantic scene reconstruction in an indoor environment with our self-developed drone.