arXiv subjects

Hamed Shah-Mansouri

Publications and source records attributed to Hamed Shah-Mansouri.

10 recordsLinked to original sources

arXiv2026

Real-Time In-Cabin Driver Behavior Recognition on Low-Cost Edge Hardware

In-cabin driver monitoring systems (DMS) must recognize distraction- and drowsiness-related behaviors with low latency under strict constraints on compute, power, and cost. We present a single-camera in-cabin driver behavior recognition system designed for deployment on two low-cost edge platforms: Raspberry Pi 5 (CPU-only) and the Google Coral development board with an Edge Tensor Processing Unit (Edge TPU) accelerator. The proposed pipeline combines (i) a compact per-frame vision model, (ii) a confounder-aware label taxonomy to reduce confusions among visually similar behaviors, and (iii) a temporal decision head that triggers alerts only when predictions are both confident and sustained. The system supports 17 behavior classes. Training and evaluation use licensed datasets plus in-house collection (over 800,000 labeled frames) with driver-disjoint splits, and we further validate the deployed system in live in-vehicle tests. End-to-end performance reaches approximately 16 FPS on Raspberry Pi 5 using 8-bit integer (INT8) inference (per-frame latency <60 ms) and approximately 25 FPS on Coral Edge TPU (end-to-end latency ~40 ms), enabling real-time monitoring and stable alert generation on embedded hardware. Finally, we discuss how reliable in-cabin perception can serve as an upstream signal for human-centered vehicle intelligence, including emerging agentic vehicle concepts.

cs.CV↗

arXiv2025

QECO: A QoE-Oriented Computation Offloading Algorithm based on Deep Reinforcement Learning for Mobile Edge Computing

In the realm of mobile edge computing (MEC), efficient computation task offloading plays a pivotal role in ensuring a seamless quality of experience (QoE) for users. Maintaining a high QoE is paramount in today's interconnected world, where users demand reliable services. This challenge stands as one of the most primary key factors contributing to handling dynamic and uncertain mobile environments. In this study, we delve into computation offloading in MEC systems, where strict task processing deadlines and energy constraints can adversely affect the system performance. We formulate the computation task offloading problem as a Markov decision process (MDP) to maximize the long-term QoE of each user individually. We propose a distributed QoE-oriented computation offloading (QECO) algorithm based on deep reinforcement learning (DRL) that empowers mobile devices to make their offloading decisions without requiring knowledge of decisions made by other devices. Through numerical studies, we evaluate the performance of QECO. Simulation results reveal that compared to the state-of-the-art existing works, QECO increases the number of completed tasks by up to 14.4%, while simultaneously reducing task delay and energy consumption by 9.2% and 6.3%, respectively. Together, these improvements result in a significant average QoE enhancement of 37.1%. This substantial improvement is achieved by accurately accounting for user dynamics and edge server workloads when making intelligent offloading decisions. This highlights QECO's effectiveness in enhancing users' experience in MEC systems.

cs.NI↗

arXiv2025

Enhanced AI as a Service at the Edge via Transformer Network

Artificial intelligence (AI) has become a pivotal force in reshaping next generation mobile networks. Edge computing holds promise in enabling AI as a service (AIaaS) for prompt decision-making by offloading deep neural network (DNN) inference tasks to the edge. However, current methodologies exhibit limitations in efficiently offloading the tasks, leading to possible resource underutilization and waste of mobile devices' energy. To tackle these issues, in this paper, we study AIaaS at the edge and propose an efficient offloading mechanism for renowned DNN architectures like ResNet and VGG16. We model the inference tasks as directed acyclic graphs and formulate a problem that aims to minimize the devices' energy consumption while adhering to their latency requirements and accounting for servers' capacity. To effectively solve this problem, we utilize a transformer DNN architecture. By training on historical data, we obtain a feasible and near-optimal solution to the problem. Our findings reveal that the proposed transformer model improves energy efficiency compared to established baseline schemes. Notably, when edge computing resources are limited, our model exhibits an 18\% reduction in energy consumption and significantly decreases task failure compared to existing works.

cs.NI↗

arXiv2024

Edge AI as a Service with Coordinated Deep Neural Networks

As artificial intelligence (AI) applications continue to expand in next-generation networks, there is a growing need for deep neural network (DNN) models. Although DNN models deployed at the edge are promising for providing AI as a service with low latency, their cooperation is yet to be explored. In this paper, we consider that DNN service providers share their computing resources as well as their models' parameters and allow other DNNs to offload their computations without mirroring. We propose a novel algorithm called coordinated DNNs on edge (\textbf{CoDE}) that facilitates coordination among DNN services by establishing new inference paths. CoDE aims to find the optimal path, which is the path with the highest possible reward, by creating multi-task DNNs from individual models. The reward reflects the inference throughput and model accuracy. With CoDE, DNN models can make new paths for inference by using their own or other models' parameters. We then evaluate the performance of CoDE through numerical experiments. The results demonstrate a $40\%$ increase in the inference throughput while degrading the average accuracy by only $2.3\%$. Experiments show that CoDE enhances the inference throughput and, achieves higher precision compared to a state-of-the-art existing method.

cs.NI↗

arXiv2023

An Intelligent Indoor Positioning Algorithm Based on Wi-Fi and Bluetooth Low Energy

Indoor positioning plays a pivotal role in a wide range of applications, from smart homes to industrial automation. In this paper, we propose a comprehensive approach for accurate positioning in indoor environments through the integration of existing Wi-Fi and Bluetooth Low Energy (BLE) devices. The proposed algorithm involves acquiring the received signal strength indicator (RSSI) data from these devices and capturing the complex interactions between RSSI and positions. To enhance the accuracy of the collected data, we first use a Kalman filter for denoising RSSI values, then categorize them into distinct classes using the K-nearest neighbor (KNN) algorithm. Incorporating the filtered RSSI data and the class information obtained from KNN, we then introduce a recurrent neural network (RNN) architecture to estimate the positions with a high precision. We further evaluate the accuracy of our proposed algorithm through testbed experiments using ESP32 system on chip with integrated Wi-Fi and BLE. The results show that we can accurately estimate the positions with an average error of 61.29 cm, which demonstrates a 56\% enhancement compared to the state-of-the-art existing works.

cs.NI↗

arXiv2017

Hierarchical Fog-Cloud Computing for IoT Systems: A Computation Offloading Game

Fog computing, which provides low-latency computing services at the network edge, is an enabler for the emerging Internet of Things (IoT) systems. In this paper, we study the allocation of fog computing resources to the IoT users in a hierarchical computing paradigm including fog and remote cloud computing services. We formulate a computation offloading game to model the competition between IoT users and allocate the limited processing power of fog nodes efficiently. Each user aims to maximize its own quality of experience (QoE), which reflects its satisfaction of using computing services in terms of the reduction in computation energy and delay. Utilizing a potential game approach, we prove the existence of a pure Nash equilibrium and provide an upper bound for the price of anarchy. Since the time complexity to reach the equilibrium increases exponentially in the number of users, we further propose a near-optimal resource allocation mechanism and prove that in a system with $N$ IoT users, it can achieve an $ε$-Nash equilibrium in $O(N/ε)$ time. Through numerical studies, we evaluate the users' QoE as well as the equilibrium efficiency. Our results reveal that by utilizing the proposed mechanism, more users benefit from computing services in comparison to an existing offloading mechanism. We further show that our proposed mechanism significantly reduces the computation delay and enables low-latency fog computing services for delay-sensitive IoT applications.

cs.DC↗

arXiv2013

Caching Gain in Wireless Networks with Fading: A Multi-User Diversity Perspective

We consider the effect of caching in wireless networks where fading is the dominant channel effect. First, we propose a one-hop transmission strategy for cache-enabled wireless networks, which is based on exploiting multi-user diversity gain. Then, we derive a closed-form result for throughput scaling of the proposed scheme in large networks, which reveals the inherent trade-off between cache memory size and network throughput. Our results show that substantial throughput improvements are achievable in networks with sources equipped with large cache size. We also verify our analytical result through simulations.

cs.IT↗

arXiv2013

Throughput of One-Hop Wireless Networks with Noisy Feedback Channel

In this paper, we consider the effect of feedback channel error on the throughput of one-hop wireless networks under the random connection model. The transmission strategy is based on activating source-destination pairs with strongest direct links. While these activated pairs are identified based on Channel State Information (CSI) at the receive side, the transmit side will be provided with a noisy version of this information via the feedback channel. Such error will degrade network throughput, as we investigate in this paper. Our results show that if the feedback error probability is below a given threshold, network can tolerate such error without any significant throughput loss. The threshold value depends on the number of nodes in the network and the channel fading distribution. Such analysis is crucial in design of error correction codes for feedback channel in such networks.

cs.IT↗

arXiv2013

Throughput of Large One-hop Wireless Networks with General Fading

Consider $n$ source-destination pairs randomly located in a shared wireless medium, resulting in interference between different transmissions. All wireless links are modeled by independently and identically distributed (i.i.d.) random variables, indicating that the dominant channel effect is the random fading phenomenon. We characterize the throughput of one-hop communication in such network. First, we present a closed-form expression for throughput scaling of a heuristic strategy, for a completely general channel power distribution. This heuristic strategy is based on activating the source-destination pairs with the best direct links, and forcing the others to be silent. Then, we present the results for several common examples, namely, Gamma (Nakagami-$m$ fading), Weibull, Pareto, and Log-normal channel power distributions. Finally -- by proposing an upper bound on throughput of all possible strategies for super-exponential distributions -- we prove that the aforementioned heuristic method is order-optimal for Nakagami-$m$ fading.

cs.IT↗

arXiv2011

One-Hop Throughput of Wireless Networks with Random Connections

We consider one-hop communication in wireless networks with random connections. In the random connection model, the channel powers between different nodes are drawn from a common distribution in an i.i.d. manner. An scheme achieving the throughput scaling of order $n^{1/3-δ}$, for any $δ>0$, is proposed, where $n$ is the number of nodes. Such achievable throughput, along with the order $n^{1/3}$ upper bound derived by Cui et al., characterizes the throughput capacity of one-hop schemes for the class of connection models with finite mean and variance.

cs.IT↗