Soft Robotic Grippers for Crop Handling or Harvesting: A Review

Abstract

Nowadays, harvesting delicate and high-value fruits, vegetables and edible fungi requires a large input of manual human labor. The relatively low wages and many health problems the workforce faces make this profession increasingly unpopular. Meanwhile, robotic systems that selectively harvest crops are being developed. Whilst the moving platform, manipulator, and recognition systems of such robots are studied the past few decades, research on the gripping end of such robots is only recently growing. This study analyses the state-of-the-art of soft grippers for crop handling and harvesting, reporting their quantitative and qualitative characteristics. Seventy-eight grippers are retrieved from the academic literature and compared with each other in terms of their design and reported performance, more specifically grasping and detachment methods, materials used, type of actuators and sensors employed, and the control of the gripping procedure. In addition, the identified grippers are classified into 13 distinct soft grasping technology categories. Moreover, the retrieved papers are analyzed with respect to their publication date and country of origin to observe the recent growth in the field. Furthermore, a subset of soft grippers is identified that was tested on the task of selectively harvesting crops, where grip and detachment success rates and plant and crop damage are compared.

Sensing soft robots' shape with cameras: an investigation on kinematics-aware SLAM

“…where tč,i i−1 and Rč,i i−1 describe the translation and rotation from the base of the segment to the camera frame according to the PCC kinematic model for an estimated configuration of the the segment qi . T č,i i−1 (q č,i ) and Rč,i i−1 (q č,i ) are based on (1) and a function of the adjusted configuration q č,i referenced in (2).…”

“…With their bodies entirely made of soft deformable materials, continuum soft robots are especially suited for application domains involving safe and robust interaction with humans and environment, and ranging from inspection, to healthcare and agriculture [1], [2]. To achieve these goals, soft robots must first master the art of controlling and sensing their body shape in space [3].…”

Agricultural Robotics and Automation [TC Spotlight]

“…Stateoftheart progress in research and future challenges is docu mented in a wide range of review papers and book chapters addressing agriculture in general [1], [3], [4], [5] and specific application domains, including phenotyping [6], [7], [8] ara ble farming [9], livestock farming [10], greenhouse horticulture [2], orchard management [11], forestry [12], and food processing [13]. Review papers also address specific technologies in the context of agricultural robotics, such as computer vision [14], [15], active perception [16], unmanned aer ial vehicle technologies [17], [18], cov erage path planning in arable farming [19], and grasping and soft grasping [20], [21]. Some illustrative examples of agrifood robotics are documented in Figure 2.…”

Soft Sensorized Physical Twin Enabling lab2field Transfer and Learning from Demonstration for Raspberry Harvesting

“…Whilst advanced machinery has enabled the harvesting of many crops from barley to beetroot, many crops have thus far escaped automation [3]. This includes those that are delicate, fragile or have highly complex surrounding environments such as berries, salad, or grapes [4], and also those which do not ripen homogeneously. Despite over 30 years of research, harvesting robots have shown limited performance improvement [5].…”

“…Whilst there have been notable examples of robotic harvesting technologies for crops including strawberries [14,15], sweet peppers [16] and apples [17], examples remain limited [18,19]. Within these, repeated periods of field trials for early stage evaluation or training are commonly reported.…”