Scaling architectures have been proven effective for improving Scene Text Recognition (STR), but the individual contribution of vision encoder and text decoder scaling remain under-explored. In this work, we present an in-depth empirical analysis and demonstrate that, contrary to previous observations, scaling the decoder yields significant performance gains, always exceeding those achieved by encoder scaling alone. We also identify label noise as a key challenge in STR, particularly in real-world data, which can limit the effectiveness of STR models. To address this, we propose Cloze Self-Distillation (CSD), a method that mitigates label noise by distilling a student model from context-aware soft predictions and pseudolabels generated by a teacher model. Additionally, we enhance the decoder architecture by introducing differential cross-attention for STR. Our methodology achieves state-of-the-art performance on 10 out of 11 benchmarks using only real data, while significantly reducing the parameter size and computational costs.

In recent years, vision transformers with text decoder have demonstrated remarkable performance on Scene Text Recognition (STR) due to their ability to capture long-range dependencies and contextual relationships with high learning capacity. However, the computational and memory demands of these models are significant, limiting their deployment in resource-constrained applications. To address this challenge, we propose an efficient and accurate STR system. Specifically, we focus on improving the efficiency of encoder models by introducing a cascaded-transformers structure. This structure progressively reduces the vision token size during the encoding step, effectively eliminating redundant tokens and reducing computational cost. Our experimental results confirm that our STR system achieves comparable performance to state-of-the-art baselines while substantially decreasing computational requirements. In particular, for large-models, the accuracy remains same, 92.77 → 92.68, while computational complexity is almost halved with our structure.

As the IoT expands exponentially, the amount of data generated by individuals has increased. To process big data efficiently, machine learning (especially deep learning) has emerged. However, existing machine learning has the disadvantage of being vulnerable to data privacy because it sends raw data to the center. Therefore, federated learning (FL) was introduced to address this privacy problem, in which only learning parameters are sent to the center after training the user’s own local model with their own raw data. However, FL remains vulnerable to various attacks. In this paper, we propose an efficient and safe FL framework using the Interplanetary File System (IPFS) that minimizes the effect of data poisoning attacks on FL. In this system, the roles of nodes are divided into three: leader node, A-node (Aggregation-node), and T-node (Training-node). In this way, the A-node and T-node cannot manipulate the learning information, allowing the sharing of information and data safely through IPFS while protecting raw data with a similarity-based data shuffling scheme used by the A-node. Moreover, nodes with high accuracy receive more incentives and learning motivation, enhancing the overall efficiency of the network. Finally, the efficiency of the system is verified through related simulations.

One of the most fundamental problems in simultaneous localization and mapping (SLAM) is the ability to take into account data association (DA) uncertainties. In this paper, this problem is addressed by proposing a multi-hypotheses sampling distribution for particle filtering-based SLAM algorithms. By modeling the measurements and landmarks as random finite sets, an importance density approximation that incorporates DA uncertainties is derived. Then, a tractable Gaussian mixture model approximation of the multi-hypotheses importance density is proposed in which each mixture component represents a different DA. Finally, an iterative method for approximating the mixture components of the sampling distribution is utilized and a partitioned update strategy is developed. Using synthetic and experimental data, it is demonstrated that the proposed importance density improves the accuracy and robustness of landmark-based SLAM in cluttered scenarios over state-of-the-art methods. At the same time, the partitioned update strategy makes it possible to include multiple DA hypotheses in the importance density approximation, leading to a favorable linear complexity scaling, in terms of the number of landmarks in the field-of-view.

Localization is envisioned as a key enabler to satisfy the requirements of communications and context-aware services in the fifth/sixth generation (5G/6G) communication systems. User localization can be achieved based on delay and angle estimation using uplink/downlink pilot signals. However, hardware impairments (HWIs) (such as phase noise and mutual coupling) distort the signals at both the transmitter and receiver sides and thus affect the localization performance. While this impact can be ignored at lower frequencies with less severe HWIs, and less stringent localization requirements, modeling and analysis efforts are needed for high-frequency bands to assess degradation in localization accuracy due to HWIs. In this work, we model various types of impairments for a mmWave multiple-input-multiple-output communication system and conduct a misspecified Cramér-Rao bound analysis to evaluate HWI-induced performance losses in terms of angle/delay estimation and the resulting 3D position/orientation estimation error. We also investigate the effect of individual and overall HWIs on communications in terms of symbol error rate (SER). Our extensive simulation results demonstrate that each type of HWI leads to a different level of degradation in angle and delay estimation performance, and the prominent impairment factors on delay estimation will have a dominant negative effect on SER.

In the upcoming sixth generation (6G) of wireless communication systems, reconfigurable intelligent surfaces (RISs) are regarded as one of the promising technological enablers, which can provide programmable signal propagation. Therefore, simultaneous radio localization and mapping (SLAM) with RISs appears as an emerging research direction within the 6G ecosystem. In this paper, we propose a novel framework of RIS-enabled radio SLAM for wireless operation without the intervention of access points (APs). We first design the RIS phase profiles leveraging prior information for the user equipment (UE), such that they uniformly illuminate the angular sector where the UE is probabilistically located. Second, we modify the marginal Poisson multi-Bernoulli SLAM filter and estimate the UE state and landmarks, which enables efficient mapping of the radio propagation environment. Third, we derive the theoretical Cramér-Rao lower bounds on the estimators for the channel parameters and the UE state. We finally evaluate the performance of the proposed method under scenarios with a limited number of transmissions, taking into account the channel coherence time. Our results demonstrate that the RIS enables solving the radio SLAM problem with zero APs, and that the consideration of the Doppler shift contributes to improving the UE speed estimates.

Millimeter-wave (mmWave) signals provide attractive opportunities for sensing due to their inherent geometrical connections to physical propagation channels. Two common modalities used in mmWave sensing are monostatic and bistatic sensing, which are usually considered separately. By integrating these two modalities, information can be shared between them, leading to improved sensing performance. In this paper, we investigate the integration of monostatic and bistatic sensing in a 5G mmWave scenario, implement the extended Kalman-Poisson multi-Bernoulli sequential filters to solve the sensing problems, and propose a method to periodically fuse user states and maps from two sensing modalities.

Radio localization is a key enabler for joint communication and sensing in the fifth/sixth generation (5G/6G) communication systems. With the help of multipath components (MPCs), localization and mapping tasks can be done with a single base station (BS) and single unsynchronized user equipment (UE) if both of them are equipped with an antenna array. However, the antenna array at the UE side increases the hardware and computational cost, preventing localization functionality. In this work, we show that with Doppler estimation and MPCs, localization and mapping tasks can be performed even with a single-antenna mobile UE. Furthermore, we show that the localization and mapping performance will improve and then saturate at a certain level with an increased UE speed. Both theoretical Cramér-Rao bound analysis and simulation results show the potential of localization under mobility and the effectiveness of the proposed localization algorithm.

Millimeter wave (mmWave) signals are useful for simultaneous localization and mapping (SLAM), due to their inherent geometric connection to the propagation environment and the propagation channel. To solve the SLAM problem, existing approaches rely on sigma-point or particle-based approximations, leading to high computational complexity, precluding real-time execution. We propose a novel low-complexity SLAM filter, based on the Poisson multi-Bernoulli mixture (PMBM) filter. It utilizes the extended Kalman (EK) first-order Taylor series based Gaussian approximation of the filtering distribution, and applies the track-oriented marginal multi-Bernoulli/Poisson (TOMB/P) algorithm to approximate the resulting PMBM as a Poisson multi-Bernoulli (PMB). The filter can account for different landmark types in radio SLAM and multiple data association hypotheses. Hence, it has an adjustable complexity/performance trade-off. Simulation results show that the developed SLAM filter can greatly reduce the computational cost, while it keeps the good performance of mapping and user state estimation.

5G new radio will provide a new paradigm in high accurate vehicle localization, reinforced by the use of large antenna arrays along with carefully designed broadband radio technology. However, a high computational load still remains unravelled in the context of cooperative localization, albeit with its advantages of high-precision localization. To alleviate such the computational burden, we develop a reliable technique of cooperative localization, addressed as a constraint satisfaction problem. A constraint satisfaction formalism for cooperative localization readily enables to recast a formulation of the decentralized optimization. Its efficient solution of the optimization is developed in a distributed manner. Simulation results demonstrate that the proposed approach saves the requested computational loads significantly while sustaining the satisfactory accuracy of cooperative localization.

5G millimeter wave (mmWave) signals can be used to jointly localize the receiver and map the propagation environment in vehicular networks, which is a typical simultaneous localization and mapping (SLAM) problem. Mapping the environment is challenging, due to measurements comprising both specular and diffuse multipath components, where diffuse multipath is usually considered as a perturbation. We here propose a novel method to utilize all available multipath signals from each landmark for mapping and incorporate this into a Poisson multi-Bernoulli mixture for the 5G SLAM problem. Simulation results demonstrate the efficacy of the proposed scheme.

In 5G mmWave, simultaneous localization and mapping (SLAM) allows devices to exploit map information to improve their position estimate. Even the most basic SLAM filter based on a Rao-Blackwellized particle filter (RBPF) combined with a probability hypothesis density (PHD) map representation exhibits high complexity. This paper proposes a new implementation method for the 5G SLAM using message passing (MP) and the cubature Kalman filter (CKF). We demonstrate that the proposed method significantly reduces the complexity while retaining the SLAM accuracy of the RBPF-PHD approach.

5G is expected to enable simultaneous vehicle localization and environment mapping (SLAM). Furthermore, vehicular networks will be covered with 5G small cells, wherein the map information is collected at each base station (BS) and then fused so as to promote the overall performance of SLAM. In 5G multi-cell SLAM, there are challenges such as the unknown number of targets, uncertainty regarding the association between the targets and the measurements, unknown types of targets, as well as map management among BSs. To address those challenges, we propose a new method for 5G multi-cell SLAM which comprises a joint cubature Kalman filter and multi-model probability hypothesis density, and a map fusion routine. Simulation results demonstrate that the proposed method solves the aforementioned challenges and also improves vehicle state and map estimates.

5G millimeter wave (mmWave) signals can enable accurate positioning in vehicular networks when the base station and vehicles are equipped with large antenna arrays. However, radio-based positioning suffers from multipath signals generated by different types of objects in the physical environment. Multipath can be turned into a benefit, by building up a radio map (comprising the number of objects, object type, and object state) and using this map to exploit all available signal paths for positioning. We propose a new method for cooperative vehicle positioning and mapping of the radio environment, comprising a multiple-model probability hypothesis density filter and a map fusion routine, which is able to consider different types of objects and different fields of views. Simulation results demonstrate the performance of the proposed method.

5G new radio (NR) provides new opportunities for accurate positioning from a single reference station: large bandwidth combined with multiple antennas, at both the base station and user sides, allows for unparalleled angle and delay resolution. Nevertheless, positioning quality is affected by multipath and clock biases. We study, in terms of performance bounds and algorithms, the ability to localize a vehicle in the presence of multipath and unknown user clock bias. We find that when a sufficient number of paths is present, a vehicle can still be localized thanks to redundancy in the geometric constraints. Moreover, the 5G NR signals enable a vehicle to build up a map of the environment.

This paper presents a cooperative localization strategy via a distributed optimization technique known as the alternating direction method of multipliers (ADMM). In a Vehicular Ad hoc Network (VANET) where a vehicle communicates with neighboring vehicles via vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication, the developed algorithm utilizes three types of measurements, which are the pairwise relative distance, angle of arrival, and absolute positions for a subset of vehicles in cooperative localization. The proposed algorithm is designed to provide an attractive solution for the localization of autonomous driving vehicle in the GPS-denied (urban) environment. Simulation results confirm the potency of distributed ADMM-based cooperative localization for autonomous driving in 5G-based VANETs.