Integrating Different Levels of Automation: Lessons From Winning the Amazon Robotics Challenge 2016

Abstract

This paper describes Team Delft's robot winning the Amazon Robotics Challenge 2016. The competition involves automating pick and place operations in semistructured environments, specifically the shelves in an Amazon warehouse. Team Delft's entry demonstrated that the current robot technology can already address most of the challenges in product handling: object recognition, grasping, motion, or task planning; under broad yet bounded conditions. The system combines an industrial robot arm, 3-D cameras and a custom gripper. The robot's software is based on the robot operating system to implement solutions based on deep learning and other state-of-the-art artificial intelligence techniques, and to integrate them with off-the-shelf components. From the experience developing the robotic system, it was concluded that: 1) the specific task conditions should guide the selection of the solution for each capability required; 2) understanding the characteristics of the individual solutions and the assumptions they embed is critical to integrate a performing system from them; and 3) this characterization can be based on "levels of robot automation." This paper proposes automation levels based on the usage of information at design or runtime to drive the robot's behavior, and uses them to discuss Team Delft's design solution and the lessons learned from this robot development experience.

Review of Deep Reinforcement Learning-Based Object Grasping: Techniques, Open Challenges, and Recommendations

“…Corbato et al [302] discussed several lessons learnt from the Amazon Robotics Challenge. The lessons include the following: task conditions must be the base guide of the solution choice, an individual solution is required for integration and problem solving should be conducted using the hierarchical structure of levels of automation [302]. Evidently, several robotic solutions for the grasping task are associated with human hands [303].…”

El voice picking: una organización del trabajo que atenta contra la subjetividad del trabajador. El caso de una cadena de comercio de la alimentación al detal

“…El sector de la gran distribución de alimentos ha sabido sacar provecho de los avances tecnológicos y ha logrado, entre otras ventajas, ahorrar enormes cantidades de costos al calcular el valor del tiempo invertido en el trabajo cotidiano (Gardes, 2018;Haase & Beimborn, 2017;Stragas & Zeimpekis, 2011;Vörös & Gemma, 2011). En los centros de distribución contemporáneos, los flujos físicos de antaño han sido reemplazados por los flujos informáticos.…”

“…Hoy, el preparador es responsable de las mismas actividades, pero obedece a un software especializado que controla todo al detalle por medio de "sistemas de reconocimiento de voz" (voice picking) (cf. Matopoulos, 2011;Haase & Beimborn, 2017) (tabla 1). Esta forma de organización del trabajo representaba, en el 2016, el 95% de los modos de preparación en las grandes plataformas logísticas del mundo (Lefeuvre-Halftermeyer et al, 2016).…”

“…Los operadores usan una pantalla montada en la cabeza (hmd) o gafas tecnológicamente modificadas que contienen una lente compatible con realidad aumentada y se superponen al mundo real con objetos virtuales (Haase & Beimborn, 2017;Guo et al, 2014).…”

“…Para algunas empresas de distribución comercial, como Amazon, por ejemplo, el voice picking está anticuado, pues han logrado la robotización total del proceso de expedición de mercancías (cf Hernández, Bharatheesha, van Egmond, Ju, & Wisse, 2018)…”