Lee Jonny, Calvo Paco
Minimal Intelligence Laboratory (MINTLab), University of Murcia, Murcia, Spain.
Front Neurorobot. 2022 Jan 31;15:772012. doi: 10.3389/fnbot.2021.772012. eCollection 2021.
Plants offer a source of bioinspiration for soft robotics. Nevertheless, a gap remains in designing robots based on the fundamental principles of plant intelligence, rooted in a non-centralized, modular architecture and a highly plastic phenotype. We contend that a holistic approach to plant bioinspiration-one that draws more fully on the features of plant intelligence and behavior-evidences the value of an enactivist perspective. This is because enactivism emphasizes not only features of embodiment such as material composition and morphology, but also autonomy as an important aspect of plant intelligence and behavior. The enactivist sense of autonomy concerns the dynamics of self-producing systems (such as plants) that create a distinction between themselves and a domain of interactions that bear on the conditions of viability of the system. This contrasts with the widespread, but diluted notion of autonomy that merely indicates the independent operability of a system for an arbitrary period. Different notions of autonomy are relevant for soft roboticists, for instance, when evaluating limitations on existing growing robots ("growbots") that take bioinspiration from plants, but depend on a fixed source of energy and material provided by an external agent. More generally, plant-inspired robots serve as a case study for an enactivist approach to intelligence, while, correspondingly, enactivism calls attention to the possibility of non-zoological forms of intelligence embodied in a self-organizing, autonomous system.
植物为软体机器人技术提供了生物灵感来源。然而,基于植物智能的基本原理设计机器人仍存在差距,这些原理植根于非集中式、模块化架构和高度可塑性的表型。我们认为,一种全面的植物生物灵感方法——更充分地借鉴植物智能和行为的特征——体现了生成认知视角的价值。这是因为生成认知不仅强调诸如物质组成和形态等具身特征,还强调自主性是植物智能和行为的一个重要方面。生成认知中的自主性概念涉及自我生成系统(如植物)的动态,这些系统在自身与影响系统生存条件的相互作用领域之间创造出区别。这与广泛但被淡化的自主性概念形成对比,后者仅仅表明系统在任意时间段内能够独立运行。不同的自主性概念对软体机器人专家很重要,例如,在评估现有的从植物获取生物灵感但依赖外部代理提供的固定能量和材料来源的生长机器人(“growbots”)的局限性时。更一般地说,受植物启发的机器人是生成认知方法研究智能的一个案例,相应地,生成认知也提请人们注意在自组织、自主系统中体现的非动物形式智能的可能性。