It is recognized that the control of mixed-autonomy platoons comprising connected and automated vehicles (CAVs) and human-driven vehicles (HDVs) can enhance traffic flow. Among existing methods, Multi-Agent Reinforcement Learning (MARL) appears to be a promising control strategy because it can manage complex scenarios in real time. However, current research on MARL-based mixed-autonomy platoon control suffers from several limitations. First, existing MARL approaches address safety by penalizing safety violations in the reward function, thus lacking theoretical safety guarantees due to the limited interpretability of RL. Second, few studies have explored the cooperative safety of multi-CAV platoons, where CAVs can be coordinated to further enhance the system-level safety involving the safety of both CAVs and HDVs. Third, existing work tends to make an unrealistic assumption that the behavior of HDVs and CAVs is publicly known and rational. To bridge the research gaps, we propose a safe MARL framework for mixed-autonomy platoons. Specifically, this framework 1) characterizes cooperative safety by designing a cooperative Control Barrier Function (CBF), enabling CAVs to collaboratively improve the safety of the entire platoon, 2) provides a safety guarantee to the MARL-based controller by integrating the CBF-based safety constraints into MARL through a differentiable quadratic programming (QP) layer, and 3) incorporates a conformal prediction module that enables each CAV to estimate the unknown behaviors of the surrounding vehicles with uncertainty qualification. Simulation results show that our proposed control strategy can effectively enhance the system-level safety through CAV cooperation of a mixed-autonomy platoon with a minimal impact on control performance.

A critical aspect of safe and efficient motion planning for autonomous vehicles (AVs) is to handle the complex and uncertain behavior of surrounding human-driven vehicles (HDVs). Despite intensive research on driver behavior prediction, existing approaches often overlook the interactions between AVs and HDVs, assuming that HDV trajectories are not influenced by AV actions. To address this gap, we present a transformer-transfer learning-based interaction-aware trajectory predictor for safe motion planning in autonomous driving, focusing on a vehicle-to-vehicle (V2V) interaction scenario involving an AV and an HDV. Specifically, we construct a transformer-based interaction-aware trajectory predictor using widely available datasets of HDV trajectory data and further transfer the learned predictor using a small set of AV-HDV interaction data. Then, to better incorporate the proposed trajectory predictor into the motion planning module of AVs, we introduce an uncertainty quantification method to characterize the predictor’s errors, which are integrated into the path-planning process. Our experimental results demonstrate the value of explicitly considering interactions and handling uncertainties.

Connected and automated vehicles (CAVs) can be beneficial for improving the operation of highway bottlenecks such as weaving sections. This paper proposes a bi-level control approach based on an upper-level deep reinforcement learning controller and a lower-level model predictive controller to coordinate the lane-changings of a mixed fleet of CAVs and human-driven vehicles (HVs) in weaving sections. The upper level represents a roadside controller that collects vehicular information from the entire weaving section and determines the control weights used in the lower-level controller. The lower level is implemented within each CAV, which takes the control weights from the upper-level controller and generates the acceleration and steering angle for individual CAVs based on the local situation. The lower-level controller further incorporates an HV trajectory predictor, which is capable of handling the dynamic topology of vehicles in weaving scenarios with intensive mandatory lane changes. The case study inspired by a real weaving section in Basel, Switzerland, shows that our method consistently outperforms state-of-the-art benchmarks.

The exponential growth of data exerts great pressure on hardware design for database systems. Memory-centric computing (MCC) architecture, which enable compute capabilities near or inside memory storage, demonstrate great potential in enhancing the efficiency of database operations with higher compute parallelism and reduced data movements. However, existing MCC architecture mainly focus on artificial intelligence (AI) computations and those designed for database applications can only run a limited number of standalone queries such as SORT or JOIN, lacking efficient support for increasingly diverse and complex database workloads. For example, realizing a commercial recommendation engine on database requires supporting workloads including but not limited to vector aggregation, convolution or$N$-hop neighborhoods computing, etc. In this work, we develop a memristor-enabled memory-centric instruction-set architecture (MeMCISA) aiming to efficiently accelerate versatile workloads in modern database systems. MeMCISA features scalable multi-bank memristor-based storage organization with near-memory circuitries and caches in banks. An out-of-order (O0O) scheduling scheme is designed for MeMCISA based on a vector instruction set with four types of instructions (bit-level, element-level, vector-level, and control-level), combining memristor-enabled in-memory computing and near-memory computing to efficiently run workloads with varying computational kernels and data sizes. MeMCISA can support parallel instruction executions across different memristor banks as well as different hardware modules within a memristor bank. Furthermore, we develop data dependency handling mechanisms to support vector dependency scenarios in MeMCISA that do not exist in conventional scalar-based instruction sets. A prototype MeMCISA is implemented based on a 40nm CMOS technology with necessary peripheral hardware including instruction buffer and instruction scheduler. To accurately study MeMCISA performance in real-world database systems, a software-hardware co-designed framework integrating reconfigurable MeMCISA prototype is created that can support end-to-end simulations for database workloads starting from raw software codes. Based on this framework, we evaluate MeMCISA performance with standalone database queries as well as complex database workloads from representative benchmarks including UniBench, neural collaborative filtering (NCF), and ResNet-18. Simulation results demonstrate that MeMCISA achieves up to 41.84 × ~ 1767.70 × in speed compared to general-purpose processors (CPUs/GPUs).

A campaign has been conducted in Qinghai Lake from July 10 to September 26, 2019. The destination of the campaign was to calibrate and validate the HY-2B and Sentinel-3A/B altimeters. A tide gauge was used to measure the water surface of the lake. A dedicated GNSS buoy was placed at the footprints of the altimeters and the tide gauge location, to measure the water surface difference between them. A static GNSS station was also placed as a reference for the GNSS buoy and then define the datum of the tide gauge. 11 radiosondes were released at the same location with the static GNSS station to achieve the wet troposphere delay (WTD). The WTD difference was 6.40 ± 5.95 mm compared with the static GNSS station derived ones. At last, the altimeters were calibrated using the tide gauge. The calibration results were + 3.38 ± 4.46 cm, −1.38 ± 4.09 cm and +2.88 ± 4.10 cm for HY-2B, Sentinel-3A and Sentinel-3B, respectively. The results showed good agreements with the cross-calibration and other calibration site.

The Sentinel-FA (S3A) altimeter flies over the Dan 'gan island (ascending Pass 309), which was located at the south of Guangdong province, China. The sea level was measured using float type water level gauge and aquatrak 5003 acoustic tide gauge to calibrate the sea surface height (SSH) of altimeters. The EGM 2008 geoid and NAO.99Jb tide model was used for geoid and tidal differences between the in-situ equipment and the footprints of the altimeter in this study. The bias of S3A altimeter was$+7.8\pm 28.9$mm in synthetic aperture radar (SAR) mode, and$+16.0\pm 38.4$mm in Low resolution mode (PLRM). The results agree well with other calibration sites. The static GNSS stations were also used for validation of microwave radiometers (MWR) in three campaigns. The wet troposphere correction derived from ECMWF and microwave radiometer product minus static GNSS stations were$+11.2\pm 11.6 mm$and$+4.4\pm 12.9$mm, with correlation of 0.980 and 0.979.

In order to accomplish the Cal/Val of altimeters more conveniently and economically, we produced a method of sea surface height measurement using an M40 unmanned surface vehicle (USV) equipped with a GPS. The accuracy of the USV in measuring the SSH was evaluated by a campaign conducted in Zhiwan island, Guangdong province. The USV was placed together with the dedicated GPS buoy behind the rear of Runjiang 1 ship, at the same time and the same location. The SSH achieved from the two equipment was then compared with each other. The results showed that: 1) the SSH measured by the USV has an equivalent accuracy with the GPS buoy, and could be used in the Cal/Val of altimeters; 2) the USV and the GPS buoy can work properly in three class sea state situation according to the SWH derived from the GPS data; 3) the attitude of the GPS buoy was much smaller than the USV, but this did not affect the accuracy of the SSH measuring; 4) the USV still have to be improved to accomplish the Cal/Val of altimeters.

A dedicated GPS bouy was used to calibrate the Sea Surface Height (SSH) of HY-2 altimeter in China sea. Six tests were conducted at Qinglan and Shidao. The results showed that: 1)HY-2A satellite altimeter has a drift of -0.021m/cycle after 41 cycle; 2)The bias of HY-2A altimeter was +3.78 ± 6.5cm using the dedicated GPS Buoy after the linear regression from 41~76 cycles; 3) The standard deviation of cross calibration was 8.75cm from 41 to 76 cycles.

The Satellite data have been more and more used in ocean forecast. They can be used not only in ocean forecast validation but also be assimilated in ocean forecast model. Since August 2011, the HY-2 satellite has been launched successfully, and now it has been used in ocean wind model. This paper is armed to use the WindSat Sea Surface Wind (SSW) data to validate the accuracy of wind forecast, which has assimilated the HY-2 Scatterometer (SCAT) SSW. Compared with HY-2 SSW, the validation results show that the accuracy of forecast wind speed improved 0.18 m/s with the accuracy of wind direction improved 0.59° after assimilating HY-2 SCAT SSW into Model, and the RE cannot accurately describe the SSW accuracy assessment.

The wet zenith delay(WZD) can be estimated by ground GPS stations as well as satellite altimeter. The Jason-2 altimeter has been an operational satellite, which provides the sea surface height with several centimeters, and have the high accuracy of measuring the WZD using a Microwave Radiometer (AMR). In this paper, we estimate the wet zenith delay (WZD) of 79 ground GPS stations using Jason-2 AMR from 2013 to 2016. The GPS stations are distributed in seven regions around the world. The distance between these GPS stations and the footprints of Jason-2 AMR was less than 150km. In the processing of GPS data, the GAMIT/GLOBK software with the VMF1 mapping function was used to achieve the GPS WZD. The results showed that the bias of the AMR and the GPS WZD was stable between the distance of 23km to 98km. The overall bias of GPS WZD was +5.68±21.96mm. Land still has a lot of influence to the AMR though new algorithm has reduced the variance within 25 km of coastline. Because of the variability of water vapor, the comparison between the GPS WZD and AMR was greatly high when the distance was more than 100km.

There are two calibration methodologies for altimeter sea surface height, tide gauge methodology and GPS buoy methodology. GPS buoy methodology need an accuracy GPS buoy which has been manufactured with GPS receiver. This article introduces in detail the assessment of the accuracy of GPS buoy by the numerical simulation and tide gauge. The simulation results show that the roll angles of manufactured GPS buoy is less than 10° and the heave rate is close to 1, which meet the challenge of altimeter sea surface height calibration on the sea experiment condition. And finally used the 4 time in-situ accuracy testing experiments, the accuracy of GPS buoy to measure water level height was accessed. The results show that the bias is 7.40cm with 0.55cm standard error. Through assessment, it can be conclude that the manufactured GPS buoy can be used in the altimeter sea surface height calibration.