Managing edge resources for fully autonomous aerial systems

Abstract

Fully autonomous aerial systems (FAAS) fly complex missions guided wholly by software. If users choose software, compute hardware and aircraft well, FAAS can complete missions faster and safer than unmanned aerial systems piloted by humans. On the other hand, poorly managed edge resources slow down missions, waste energy and inflate costs. This paper presents a model-driven approach to manage FAAS. We fly real FAAS missions, profile compute and aircraft resource usage and model expected demands. Naive profiling approaches use traces from previous flights to infer resource usage. However, edge resources can affect where FAAS fly and which data they sense. Usage profiles can diverge greatly across edge management policies. Instead of using traces, we characterize whole flight areas to accurately model resource usage for any flight path. We combine expected resource demands to model mission throughput, i.e., missions completed per fully charged battery. We validated our model by creating FAAS, measuring mission throughput across many system settings. Our FAAS benchmarks, released through our open source FAAS suite SoftwarePilot, execute realistic missions: autonomous photography, search and rescue, and agricultural scouting using well-known software. Our model predicted throughput with 4% error across mission, software and hardware settings. Competing approaches yielded 10-24% error. We used our SoftwarePilot benchmarks to study (1) GPU acceleration, scale up, and scale out, (2) onboard, edge and cloud computing, (3) energy and monetary budgets, and (4) software driven GPU management. We found that model-driven management can boost mission throughput by 10X and reduce costs by 87%.

Whole-Field Reinforcement Learning: A Fully Autonomous Aerial Scouting Method for Precision Agriculture

“…Here, utility gained after taking an action is defined as the improvement in crop health map accuracy. We use an unsupervised learning approach from prior work [ 17 ] based on k-nearest neighbors to determine the most similar prior observations from our dataset. We then compute the normalized utility gain garnered from taking each available flight action in our K similar observations.…”

“…These activities also delay missions. It can take a full 8-hours workday to exhaustively collect high definition images from every zone in an 80-acre crop field [ 11 , 17 ]. Thus, for UAS with onboard IoT systems, it is crucial to collect as much information as possible within a given UAS flight.…”

“…For example, Lin et al relied on narrowly defined tasks to trace and model compute needs for autonomous cars [ 18 ]. Boubin et al broadened Lin’s compute modeling by capturing environmental factors for UAS [ 17 ]. In-situ AI [ 19 ] and Boroujerdian et al [ 13 ] generalized these approaches via environmental simulation.…”

Latency-aware Resource Management at the Edge

“…Ahvar [67] Aït-Salaht [85] Bittencourt [78] Borgia [86] Borylo [71] Boubin [87] Chen [81] De Maio [88] Elgazar [83] Fahs [89] Fricker [72] Gu [80] Habak [20] Liu [77] Liu [50] Mahmud [90] Penner [52] Rodrigues [73] Sardellitti [82] Singh [40] Skarlat [46] Sonmez [91] Tang [79] Tärneberg [75] Wang [45] Wang [70] Wang [92] Wang [93] Xia [94] Yi [76] Zamani [74] 3.4. Objective user.…”

“…Focusing on energy, Ahvar et al [67] present a model for estimating the energy consumption of different edge and cloud architectures. Similarly, Boubin et al [87] describe a model for estimating how a certain hardware/software configuration will impact mission throughput.…”