By utilizing the mapping-based pulse width modulation (MPWM), the look-up-table (LUT) based all digital transmitter (ADT) features a superb in-band signal-to-noise ratio (SNR), but its rather high random-access memory (RAM) consumption and the computation complexity hinder its further usage in applications requiring real-time reconfigurability. Our previous work leverages a Viterbi-based PWM (VbPWM) scheme to provide an iterative deduction approach, but the degraded performance and the excessive register usage make it unsuitable for real applications. By exploring the quais-periodicity inside the differenced target function, a Viterbi-based multi-bit deduction PWM (VMbPWM) is proposed in this paper, which leverages the periodic-like feature of the target function to form a multiple-bit deduction process of the 1-bit pulse sequence. Numerical simulations suggest that the proposed scheme outperforms the MPWM by a maximum of 19 dB in in-band SNR when deducing 8 bits in a single iteration, and a maximum of 10 dB increase in SNR is presented for the VMbPWM over VbPWM when the two possess the same computation complexity. After combining iterations that deduce different amounts of bits, the computation complexity of VMbPWM can be reduced to nearly 1/4 of that when deducing 8 bits in each iteration at the cost of a 5 dB SNR degradation. Accordingly, a fully pipelined ADT structure is conducted based on the VMbPWM and is implemented with the ZCU102 evaluation kit. As depicted by the implementation results, the proposed VMbPWM-based ADT reduces the RAM usage to two tiles while eliminating the register usage from 22% using VbPWM to 15% using VMbPWM. Moreover, a reduced error vector magnitude (EVM) is demonstrated for the VMbPWM, where the maximum EVM turns out to be 2.6%, 4.0%, and 4.9% under the symbol rate of 5 MBaud, 25 MBaud, and 50 MBaud using 64QAM, which is typically 0.1%-0.2% lower than the VbPWM.

Based on the concept of mapping-based pulse width modulation (MPWM), the look-up-table-based all-digital transmitter (LUT-based ADT) achieves high signal-to-noise ratio (SNR) and low error vector magnitude (EVM) by leveraging a stored mapping from the baseband complex sample to the output binary sequence, while its high RAM consumption and LUT generation complexity hinder its further usage. Toward a RAM-free and traversal-free ADT based on MPWM, in this paper, we model the MPWM process as an optimization problem and correspondingly construct a Viterbi-based PWM (VbPWM) modulator scheme, featuring real-time deduction of the 1-bit sequence from the input complex sample and a traversal-free computation process. Fixed points simulations suggest that the deduced local optimum demonstrates a 10 dB SNR loss compared with the global optimum of MPWM, while this loss can be further eased by 2-3 dB after introducing a simplified simulated annealing algorithm. Additionally, a fully serial and pipelined implementation structure is proposed to unfold the iterations into cascading iteration units, thereby achieving run-time reconfigurable carrier frequencies and minimizing block read-only memory (BRAM) usage. EVMs of the proposed ADT with carrier frequencies ranging from 0.6 GHz and 7.6 GHz under different symbol rates are measured with ZCU102, presenting an increasing trend associated with carrier frequency and symbol rates. Consequently, compared with the original MPWM, a 1% loss of EVM is revealed under 5 MBaud of symbol rate for all measured frequencies. The maximum EVM under 5 MBaud is measured as 3.5%, while the maximum for 25 MBaud and 50 MBaud is 5% and 5.8%, respectively. The RAM usage is then minimized to a single 36K tile and the traversal of MPWM is removed.

Monocular 3D lane detection plays a crucial role in autonomous driving, assisting vehicles in safe navigation. Existing methods primarily utilize calibrated camera parameters in the dataset to conduct 3D lane detection from a single image. However, errors or sudden absence of camera parameters can pose significant challenges to safe driving. On one hand, this can lead to incorrect feature acquisition, which further affects the precision of lane detection. On the other hand, it renders methods relying on transformation matrices for temporal fusion ineffective. To address the above issue and achieve accurate 3D lane detection, we propose CaliFree3DLane, a calibration-free method for spatio-temporal 3D lane detection based on Transformer structure. Instead of using geometric projections to obtain static reference points on images, we propose a reference point refinement strategy that dynamically updates the reference points and finally generates appropriate sampling points for image feature extraction. To integrate multi-frame features, we generate sub-queries from the current scene query to focus on the image features of each frame independently. We then aggregate these sub-queries to form a more comprehensive scene query for 3D lane detection. Using these operations, CaliFree3DLane accurately transforms multi-frame image features into the current bird’s-eye view (BEV) space, enabling precise 3D lane detection. Experimental results show that our CaliFree3DLane achieves state-of-the-art 3D lane detection performance in various datasets. Compared to the Transformer-based methods of the same type, we have also improved$ {6.0\%}\sim {10.5\%}$at the F1 score. Code is available athttps://github.com/Ciisrlab/CaliFree3DLane.

The image enhancement task requires a complex balance between extracting high-level contextual information and optimizing spatial details in the image to improve the visual quality. Most of existing methods have limited capability in capturing contextual features and optimizing spatial details when they only rely on a single modality. To address the above issues, this paper introduces a novel multi-modal image enhancement network based on Transformer, named as TITFormer, which combines textual and simulating infrared modalities firstly for this important task. TITFormer comprises a text channel attention fusion (TCF) network block and an infrared-guided spatial detail optimization (SDO) network block. The TCF extracts contextual features from the high-dimensional features compressed after spatial channel transformation of the textual feature and image feature. The SDO module uses simulating infrared images characterized by pixel intensity to guide the optimization of spatial details with contextual features adaptively. Experimental results demonstrate that TITFormer achieves state-of-the-art performance on two publicly available benchmark datasets.

Researchers have applied 3D Lookup Tables (LUTs) in cameras, offering new possibilities for enhancing image quality and achieving various tonal effects. However, these approaches often overlook the non-uniformity of color distribution in the original images, which limits the performance of learnable LUTs. To address this issue, we introduce a lightweight end-to-end image enhancement method called Simulated Infrared Fusion Guided Image-adaptive 3D Lookup Tables (SIRLUT). SIRLUT enhances the adaptability of 3D LUTs by reorganizing the color distribution of images through the integration of simulated infrared imagery. Specifically, SIRLUT consists of an efficient Simulated Infrared Fusion (SIF) module and a Simulated Infrared Guided (SIG) refinement module. The SIF module leverages a cross-modal channel attention mechanism to perceive global information and generate dynamic 3D LUTs, while the SIG refinement module blends simulated infrared images to match image consistency features from both structural and color aspects, achieving local feature fusion. Experimental results demonstrate that SIRLUT outperforms state-of-the-art methods on different tasks by up to 0.88 ~ 2.25dB while reducing the number of parameters. Code is available at https://github.com/riversky2025/SIRLUT.git .

The bandwidth of the frequency-modulated continuous wave (FMCW) system is largely increased with the progress in high-speed technology to achieve high accuracy ranging. In situations where bandwidth is restricted, the range estimation algorithm is still worth concern. By breaking down the range estimation of the FMCW system into single-tone frequency estimation, in this paper we propose a Phase-Difference-based Single-Tone estimation (PDb-STE) algorithm consisting of two fast Fourier transform (FFT) units and an interpolator. Spectra obtained from the FFT units are utilized to form a phase difference, which is used for the FFT enhancement. Frequency, phase and amplitude estimations are carried out based on the corrected phase difference. Theoretical proof of the proposed algorithm is given under the rectangular window and further applied with Hann, Hanning, Blackman and Kaiser windows during simulation. The results show that for Kaiser window the expanded main lobe brings less systematic error for both complex signal estimation and in-phase part estimation. Simulation of FMCW system is performed based on proposed algorithm. By setting iteration chirp-Z transform (CZT) as the comparison, the proposed algorithm outperforms the CZT by 1 dB under rectangular and Kaiser window with the main lobe control factor β ≤ 2.5. The performance worsens as the β grows, which is due to the increased noise disruption to the wide main lobe. Hardware verification is carried out to show the feasibility of the algorithm and obtain results with all root mean square error(RMSE) below 0.02 m.