Showing 1–4 of 4 results for author: Pereira, P G

Reducing FEC-Complexity in Cross-Layer Predictable Data Communication

Authors: Pablo Gil Pereira, Thorsten Herfet

Abstract: The PRRT protocol enables applications with strict performance requirements such as Cyber-Physical Systems, as it provides predictably low, end-to-end delay via cross-layer pacing and timely error correction via Hybrid ARQ (HARQ). However, the implemented HARQ uses computationally complex Maximum Distance Separable (MDS) codes to generate redundancy. In this paper we propose code partitioning for… ▽ More The PRRT protocol enables applications with strict performance requirements such as Cyber-Physical Systems, as it provides predictably low, end-to-end delay via cross-layer pacing and timely error correction via Hybrid ARQ (HARQ). However, the implemented HARQ uses computationally complex Maximum Distance Separable (MDS) codes to generate redundancy. In this paper we propose code partitioning for the complexity reduction of MDS codes, thereby enabling their deployment on constrained embedded devices. △ Less

Submitted 21 February, 2022; originally announced February 2022.

Polar Coding for Efficient Transport Layer Multicast

Authors: Pablo Gil Pereira, Thorsten Herfet

Abstract: In this paper, we shed light on how an adaptive, efficient error coding in the transport layer helps ensure the application requirements. We recap the use of MDS codes and show that binary coding can significantly reduce the complexity and hence increase the applicability also for embedded devices. We exploit the persymmetric structure of the generator matrix in polar codes to establish a duality… ▽ More In this paper, we shed light on how an adaptive, efficient error coding in the transport layer helps ensure the application requirements. We recap the use of MDS codes and show that binary coding can significantly reduce the complexity and hence increase the applicability also for embedded devices. We exploit the persymmetric structure of the generator matrix in polar codes to establish a duality of dispersion over channels (the polarization effect) and over packets (the generality required for multicast transmission), thereby constructing systematic polar codes for incremental redundancy whose performance, despite a much lower complexity, is near to MDS codes for medium-range residual loss rates. △ Less

Submitted 21 February, 2022; originally announced February 2022.

Latency-aware and -predictable Communication with Open Protocol Stacks for Remote Drone Control

Authors: Marlene Böhmer, Andreas Schmidt, Pablo Gil Pereira, Thorsten Herfet

Abstract: In order to create cooperating swarms of Unmanned Autonomous Vehicles (UAVs) that also interact with various other systems and devices, open and free communication systems are mandatory. This paper presents an implementation of such a communication system to incorporate the Crazyflie nano-drone as a UAV platform. The protocol stack leverages the open Predictably Reliable Real-time Transport (PRRT)… ▽ More In order to create cooperating swarms of Unmanned Autonomous Vehicles (UAVs) that also interact with various other systems and devices, open and free communication systems are mandatory. This paper presents an implementation of such a communication system to incorporate the Crazyflie nano-drone as a UAV platform. The protocol stack leverages the open Predictably Reliable Real-time Transport (PRRT) protocol that adds latency-awareness and -predictability to stacks composed of standard Internet protocols. To enable the drone to receive and reply to control commands via Wi-Fi, it has been extended with a Raspberry Pi that runs two variants of the Crazybridge---a software to connect the control board to the network. To evaluate how practical this solution is for the use in control applications, the communication has been analysed with a focus on the latency properties. Our investigations show that despite using the open protocol stack---and hence opting out of specialised implementations---the resulting latencies are in the same order of magnitude (4 to 9 ms) as the latency of the proprietary link. △ Less

Submitted 19 October, 2020; originally announced October 2020.

Cross-Layer Effects on Training Neural Algorithms for Video Streaming

Authors: Pablo Gil Pereira, Andreas Schmidt, Thorsten Herfet

Abstract: Nowadays Dynamic Adaptive Streaming over HTTP (DASH) is the most prevalent solution on the Internet for multimedia streaming and responsible for the majority of global traffic. DASH uses adaptive bit rate (ABR) algorithms, which select the video quality considering performance metrics such as throughput and playout buffer level. Pensieve is a system that allows to train ABR algorithms using reinfo… ▽ More Nowadays Dynamic Adaptive Streaming over HTTP (DASH) is the most prevalent solution on the Internet for multimedia streaming and responsible for the majority of global traffic. DASH uses adaptive bit rate (ABR) algorithms, which select the video quality considering performance metrics such as throughput and playout buffer level. Pensieve is a system that allows to train ABR algorithms using reinforcement learning within a simulated network environment and is outperforming existing approaches in terms of achieved performance. In this paper, we demonstrate that the performance of the trained ABR algorithms depends on the implementation of the simulated environment used to train the neural network. We also show that the used congestion control algorithm impacts the algorithms' performance due to cross-layer effects. △ Less

Submitted 27 September, 2018; originally announced September 2018.