arXiv subjects

Mengmeng Liu

Publications and source records attributed to Mengmeng Liu.

34 records · Page 2Linked to original sources

arXiv2023

Tracing the Influence of Predecessors on Trajectory Prediction

In real-world traffic scenarios, agents such as pedestrians and car drivers often observe neighboring agents who exhibit similar behavior as examples and then mimic their actions to some extent in their own behavior. This information can serve as prior knowledge for trajectory prediction, which is unfortunately largely overlooked in current trajectory prediction models. This paper introduces a novel Predecessor-and-Successor (PnS) method that incorporates a predecessor tracing module to model the influence of predecessors (identified from concurrent neighboring agents) on the successor (target agent) within the same scene. The method utilizes the moving patterns of these predecessors to guide the predictor in trajectory prediction. PnS effectively aligns the motion encodings of the successor with multiple potential predecessors in a probabilistic manner, facilitating the decoding process. We demonstrate the effectiveness of PnS by integrating it into a graph-based predictor for pedestrian trajectory prediction on the ETH/UCY datasets, resulting in a new state-of-the-art performance. Furthermore, we replace the HD map-based scene-context module with our PnS method in a transformer-based predictor for vehicle trajectory prediction on the nuScenes dataset, showing that the predictor maintains good prediction performance even without relying on any map information.

cs.RO↗

arXiv2023

An End-to-End Framework of Road User Detection, Tracking, and Prediction from Monocular Images

Perception that involves multi-object detection and tracking, and trajectory prediction are two major tasks of autonomous driving. However, they are currently mostly studied separately, which results in most trajectory prediction modules being developed based on ground truth trajectories without taking into account that trajectories extracted from the detection and tracking modules in real-world scenarios are noisy. These noisy trajectories can have a significant impact on the performance of the trajectory predictor and can lead to serious prediction errors. In this paper, we build an end-to-end framework for detection, tracking, and trajectory prediction called ODTP (Online Detection, Tracking and Prediction). It adopts the state-of-the-art online multi-object tracking model, QD-3DT, for perception and trains the trajectory predictor, DCENet++, directly based on the detection results without purely relying on ground truth trajectories. We evaluate the performance of ODTP on the widely used nuScenes dataset for autonomous driving. Extensive experiments show that ODPT achieves high performance end-to-end trajectory prediction. DCENet++, with the enhanced dynamic maps, predicts more accurate trajectories than its base model. It is also more robust when compared with other generative and deterministic trajectory prediction models trained on noisy detection results.

cs.CV↗

arXiv2023

Reduced-Complexity Cross-Domain Iterative Detection for OTFS Modulation via Delay-Doppler Decoupling

In this paper, a reduced-complexity cross-domain iterative detection for orthogonal time frequency space (OTFS) modulation is proposed, which exploits channel properties in both time and delay-Doppler domains. Specifically, we first show that in the time domain effective channel, the path delay only introduces interference among samples in adjacent time slots, while the Doppler becomes a phase term that does not affect the channel sparsity. This ``band-limited'' matrix structure motivates us to apply a reduced-size linear minimum mean square error (LMMSE) filter to eliminate the effect of delay in the time domain, while exploiting the cross-domain iteration for minimizing the effect of Doppler by noticing that the time and Doppler are a pair of Fourier dual. The state (MSE) evolution was derived and compared with bounds to verify the effectiveness of the proposed scheme. Simulation results demonstrate that the proposed scheme achieves almost the same error performance as the optimal detection, but only requires a reduced complexity.

cs.IT↗

arXiv2023

GATraj: A Graph- and Attention-based Multi-Agent Trajectory Prediction Model

Trajectory prediction has been a long-standing problem in intelligent systems like autonomous driving and robot navigation. Models trained on large-scale benchmarks have made significant progress in improving prediction accuracy. However, the importance on efficiency for real-time applications has been less emphasized. This paper proposes an attention-based graph model, named GATraj, which achieves a good balance of prediction accuracy and inference speed. We use attention mechanisms to model the spatial-temporal dynamics of agents, such as pedestrians or vehicles, and a graph convolutional network to model their interactions. Additionally, a Laplacian mixture decoder is implemented to mitigate mode collapse and generate diverse multimodal predictions for each agent. GATraj achieves state-of-the-art prediction performance at a much higher speed when tested on the ETH/UCY datasets for pedestrian trajectories, and good performance at about 100 Hz inference speed when tested on the nuScenes dataset for autonomous driving. We conduct extensive experiments to analyze the probability estimation of the Laplacian mixture decoder and compare it with a Gaussian mixture decoder for predicting different multimodalities. Furthermore, comprehensive ablation studies demonstrate the effectiveness of each proposed module in GATraj. The code is released at https://github.com/mengmengliu1998/GATraj.

cs.RO↗

arXiv2023

LAformer: Trajectory Prediction for Autonomous Driving with Lane-Aware Scene Constraints

Trajectory prediction for autonomous driving must continuously reason the motion stochasticity of road agents and comply with scene constraints. Existing methods typically rely on one-stage trajectory prediction models, which condition future trajectories on observed trajectories combined with fused scene information. However, they often struggle with complex scene constraints, such as those encountered at intersections. To this end, we present a novel method, called LAformer. It uses a temporally dense lane-aware estimation module to select only the top highly potential lane segments in an HD map, which effectively and continuously aligns motion dynamics with scene information, reducing the representation requirements for the subsequent attention-based decoder by filtering out irrelevant lane segments. Additionally, unlike one-stage prediction models, LAformer utilizes predictions from the first stage as anchor trajectories and adds a second-stage motion refinement module to further explore temporal consistency across the complete time horizon. Extensive experiments on Argoverse 1 and nuScenes demonstrate that LAformer achieves excellent performance for multimodal trajectory prediction.

cs.CV↗

arXiv2022

Unsupervised Domain Adaptive Salient Object Detection Through Uncertainty-Aware Pseudo-Label Learning

Recent advances in deep learning significantly boost the performance of salient object detection (SOD) at the expense of labeling larger-scale per-pixel annotations. To relieve the burden of labor-intensive labeling, deep unsupervised SOD methods have been proposed to exploit noisy labels generated by handcrafted saliency methods. However, it is still difficult to learn accurate saliency details from rough noisy labels. In this paper, we propose to learn saliency from synthetic but clean labels, which naturally has higher pixel-labeling quality without the effort of manual annotations. Specifically, we first construct a novel synthetic SOD dataset by a simple copy-paste strategy. Considering the large appearance differences between the synthetic and real-world scenarios, directly training with synthetic data will lead to performance degradation on real-world scenarios. To mitigate this problem, we propose a novel unsupervised domain adaptive SOD method to adapt between these two domains by uncertainty-aware self-training. Experimental results show that our proposed method outperforms the existing state-of-the-art deep unsupervised SOD methods on several benchmark datasets, and is even comparable to fully-supervised ones.

cs.CV↗

arXiv2017

Monte Carlo Study of Temperature-dependent Non-diffusive Thermal Transport in Si Nanowires

Non-diffusive thermal transport has gained extensive research interest recently due to its important implications on fundamental understanding of material phonon mean free path distributions and many nanoscale energy applications. In this work, we systematically investigate the role of boundary scattering and nanowire length on the nondiffusive thermal transport in thin silicon nanowires by rigorously solving the phonon Boltzmann transport equation using a variance reduced Monte Carlo technique across a range of temperatures. The simulations use the complete phonon dispersion and spectral lifetime data obtained from first-principle density function theory calculations as input without any adjustable parameters. Our BTE simulation results show that the nanowire length plays an important role in determining the thermal conductivity of silicon nanowires. In addition, our simulation results suggest significant phonon confinement effect for the previously measured silicon nanowires. These findings are important for a comprehensive understanding of microscopic non-diffusive thermal transport in silicon nanowires.

physics.comp-ph↗

arXiv2017

Monte Carlo Simulation of Single-Crystalline PbSe Nanowire Thermal Conductivity Using First-Principle Phonon Properties

Prior experimental studies showed that nanowires are promising structures for improving the thermoelectric performance of practical thermoelectric materials due to the strongly induced phonon-boundary scattering. However, few studies examined the impact of phonon-boundary scattering on the thermal conductivity of thermoelectric nanowires from a first-principle approach. In this work, we systematically study the role of phonon-boundary scattering with different boundary specularities on the thermal conductivity of PbSe nanowires by rigorously solving the full phonon Boltzmann transport equation without any adjustable parameters. We observe significant thermal conductivity reduction for rough PbSe nanowires with diameters less than a few hundred nanometers. The reduction reaches ~ 40% for 10 nm thick rough PbSe nanowires at room temperature. The diameter-dependent thermal conductivities are found to contain important information about the phonon mean free path distribution from a standard reconstruction algorithm. The simulation results are important for fundamental understanding of nanoscale thermal transport in thermoelectric materials and will guide future design of thermoelectric devices to achieve better energy conversion efficiency.

cond-mat.mes-hall↗

arXiv2014

Enabling Incremental Query Re-Optimization

As declarative query processing techniques expand in scope --- to the Web, data streams, network routers, and cloud platforms --- there is an increasing need for adaptive query processing techniques that can re-plan in the presence of failures or unanticipated performance changes. A status update on the data distributions or the compute nodes may have significant repercussions on the choice of which query plan should be running. Ideally, new system architectures would be able to make cost-based decisions about reallocating work, migrating data, etc., and react quickly as real-time status information becomes available. Existing cost-based query optimizers are not incremental in nature, and must be run "from scratch" upon each status or cost update. Hence, they generally result in adaptive schemes that can only react slowly to updates. An open question has been whether it is possible to build a cost-based re-optimization architecture for adaptive query processing in a streaming or repeated query execution environment, e.g., by incrementally updating optimizer state given new cost information. We show that this can be achieved beneficially, especially for stream processing workloads. Our techniques build upon the recently proposed approach of formulating query plan enumeration as a set of recursive datalog queries; we develop a variety of novel optimization approaches to ensure effective pruning in both static and incremental cases. We implement our solution within an existing research query processing system, and show that it effectively supports cost-based initial optimization as well as frequent adaptivity.

cs.DB↗

arXiv2013

Tricyclic graphs with maximal revised Szeged index

The revised Szeged index of a graph $G$ is defined as $Sz^*(G)=\sum_{e=uv \in E}(n_u(e)+ n_0(e)/2)(n_v(e)+ n_0(e)/2),$ where $n_u(e)$ and $n_v(e)$ are, respectively, the number of vertices of $G$ lying closer to vertex $u$ than to vertex $v$ and the number of vertices of $G$ lying closer to vertex $v$ than to vertex $u$, and $n_0(e)$ is the number of vertices equidistant to $u$ and $v$. In this paper, we give an upper bound of the revised Szeged index for a connected tricyclic graph, and also characterize those graphs that achieve the upper bound.

math.CO↗

arXiv2013

Further hardness results on the generalized connectivity of graphs

The generalized $k$-connectivity $κ_k(G)$ of a graph $G$ was introduced by Chartrand et al. in 1984, which is a nice generalization of the classical connectivity. Recently, as a natural counterpart, Li et al. proposed the concept of generalized edge-connectivity for a graph. In this paper, we determine the computational complexity of the generalized connectivity and generalized edge-connectivity of a graph. Two conjectures are also proved to be true.

math.CO↗

arXiv2012

The (revised) Szeged index and the Wiener index of a nonbipartite graph

Hansen et. al. used the computer programm AutoGraphiX to study the differences between the Szeged index $Sz(G)$ and the Wiener index $W(G)$, and between the revised Szeged index $Sz^*(G)$ and the Wiener index for a connected graph $G$. They conjectured that for a connected nonbipartite graph $G$ with $n \geq 5$ vertices and girth $g \geq 5,$ $ Sz(G)-W(G) \geq 2n-5. $ Moreover, the bound is best possible as shown by the graph composed of a cycle on 5 vertices, $C_5$, and a tree $T$ on $n-4$ vertices sharing a single vertex. They also conjectured that for a connected nonbipartite graph $G$ with $n \geq 4$ vertices, $ Sz^*(G)-W(G) \geq \frac{n^2+4n-6}{4}. $ Moreover, the bound is best possible as shown by the graph composed of a cycle on 3 vertices, $C_3$, and a tree $T$ on $n-3$ vertices sharing a single vertex. In this paper, we not only give confirmative proofs to these two conjectures but also characterize those graphs that achieve the two lower bounds.

math.CO↗

arXiv2012

On a relation between the Szeged index and the Wiener index for bipartite graphs

{\small The Wiener index $W(G)$ of a graph $G$ is the sum of the distances between all pairs of vertices in the graph. The Szeged index $Sz(G)$ of a graph $G$ is defined as $Sz(G)=\sum_{e=uv \in E}n_u(e)n_v(e)$ where $n_u(e)$ and $n_v(e)$ are, respectively, the number of vertices of $G$ lying closer to vertex $u$ than to vertex $v$ and the number of vertices of $G$ lying closer to vertex $v$ than to vertex $u$. Hansen used the computer programm AutoGraphiX and made the following conjecture about the Szeged index and the Wiener index for a bipartite connected graph $G$ with $n \geq 4$ vertices and $m \geq n$ edges: $$ Sz(G)-W(G) \geq 4n-8. $$ Moreover the bound is best possible as shown by the graph composed of a cycle on 4 vertices $C_4$ and a tree $T$ on $n-3$ vertices sharing a single vertex. This paper is to give a confirmative proof to this conjecture.

math.CO↗

arXiv2011

Rainbow connection number of dense graphs

An edge-colored graph $G$ is rainbow connected, if any two vertices are connected by a path whose edges have distinct colors. The rainbow connection number of a connected graph $G$, denoted $rc(G)$, is the smallest number of colors that are needed in order to make $G$ rainbow connected. In this paper we show that $rc(G)\leq 3$, if $|E(G)|\geq {{n-2}\choose 2}+2$, and $rc(G)\leq 4$, if $|E(G)|\geq {{n-3}\choose 2}+3$. These bounds are sharp.

math.CO↗

arXiv2011

Bicyclic graphs with maximal revised Szeged index

The revised Szeged index $Sz^*(G)$ is defined as $Sz^*(G)=\sum_{e=uv \in E}(n_u(e)+ n_0(e)/2)(n_v(e)+ n_0(e)/2),$ where $n_u(e)$ and $n_v(e)$ are, respectively, the number of vertices of $G$ lying closer to vertex $u$ than to vertex $v$ and the number of vertices of $G$ lying closer to vertex $v$ than to vertex $u$, and $n_0(e)$ is the number of vertices equidistant to $u$ and $v$. Hansen used the AutoGraphiX and made the following conjecture about the revised Szeged index for a connected bicyclic graph $G$ of order $n \geq 6$: $$ Sz^*(G)\leq \{{array}{ll} (n^3+n^2-n-1)/4,& {if $n$ is odd}, (n^3+n^2-n)/4, & {if $n$ is even}. {array}. $$ with equality if and only if $G$ is the graph obtained from the cycle $C_{n-1}$ by duplicating a single vertex. This paper is to give a confirmative proof to this conjecture.

math.CO↗

arXiv2011

Nordhaus-Gaddum-type theorem for the rainbow vertex-connection number of a graph

A vertex-colored graph $G$ is rainbow vertex-connected if any pair of distinct vertices are connected by a path whose internal vertices have distinct colors. The rainbow vertex-connection number of $G$, denoted by $rvc(G)$, is the minimum number of colors that are needed to make $G$ rainbow vertex-connected. In this paper we give a Nordhaus-Gaddum-type result of the rainbow vertex-connection number. We prove that when $G$ and $\bar{G}$ are both connected, then $2\leq rvc(G)+rvc(\bar{G})\leq n-1$. Examples are given to show that both the upper bound and the lower bound are best possible for all $n\geq 5$.

math.CO↗