AutoSoC: Automating Algorithm-SOC Co-design for Aerial Robots

Abstract

Aerial autonomous machines (Drones) has a plethora of promising applications and use cases. While the popularity of these autonomous machines continues to grow, there are many challenges, such as endurance and agility, that could hinder the practical deployment of these machines. The closed-loop control frequency must be high to achieve high agility. However, given the resource-constrained nature of the aerial robot, achieving high control loop frequency is hugely challenging and requires careful co-design of algorithm and onboard computer. Such an effort requires infrastructures that bridge various domains, namely robotics, machine learning, and system architecture design. To that end, we present AutoSoC, a framework for co-designing algorithms as well as hardware accelerator systems for end-to-end learning-based aerial autonomous machines. We demonstrate the efficacy of the framework by training an obstacle avoidance algorithm for aerial robots to navigate in a densely cluttered environment. For the best performing algorithm, our framework generates various accelerator design candidates with varying performance, area, and power consumption. The framework also runs the ASIC flow of place and route and generates a layout of the floor-planed accelerator, which can be used to tape-out the final hardware chip.

Analyzing and Improving Fault Tolerance of Learning-Based Navigation Systems

“…7e demonstrates the fault tolerance of three fixed-point data types: Q (1,4,11), Q (1,7,8) and Q (1,10,5) (Q (sign, integer, fraction)) with the same flight quality baseline. We observe that Q (1,4,11) consistently exhibit higher resilience than Q (1,7,8) and Q (1,10,5). This is because Q (1,4,11) has a narrower range compared to others, and are already able to capture the range of weights without performance loss.…”

“…Autonomous navigation continues to be deployed and attains significant traction in the field of robotics, unmanned drones and autonomous vehicles at all computing scales [1][2][3][4][5][6]. It helps an agent avoid navigating in unknown environments and situations; and take safe actions as needed.…”

Heterogeneous Flight Management System (FMS) Design for Unmanned Aerial Vehicles (UAVs): Current Stages, Challenges, and Opportunities

“…Given the resource-constrained nature of the UAV, achieving high control and managing frequency presents a problematic task [122]. Furthermore, prior studies have shown that computing latency is essential for increasing the computing speed of UAV, which leads to low-power consumption during missing process [123].…”

Circuit and System Technologies for Energy-Efficient Edge Robotics: (Invited Paper)

“…MAVFI [41] proposes a fault injection framework for end-to-end reliability characterization of robotic workload, which is portable to robot operating system (ROS)-based applications. Based on AirLearning, Skyline [42], Autopilot [43], and AutoSoC [44] propose a visual performance model and automated design space exploration framework to design optimal onboard compute for aerial robots. Based on PEDRA, Anwar et al [45] present a transfer learning-based approach to reduce the onboard computation required to train a neural network for autonomous navigation.…”