.png){kind=icon}
.png){kind=icon}
Return
Integrated in-memory and near-memory sensor for bionic robot perception
.png){kind=icon}
PRE
AI
Q
X
T
J
P
Q
C
J
DOI：10.1016/j.mattod.2026.103345.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The rapid development of sensor networks and edge intelligence has intensified challenges in energy-efficient data acquisition, transmission, and processing. Conventional von Neumann architectures prove inadequate for handling massive distributed sensory data, exacerbating energy consumption. To address these limitations, integrated sensing-storage-computing strategies are emerging as transformative solutions. These are primarily realized through the in-memory sensor, where the memory device itself intrinsically senses stimuli, and the near-memory sensor, which couples discrete sensor and memory components. Mirroring human multisensory integration, artificial systems aim to unify environmental awareness through coordinated visual, tactile, olfactory, auditory, and gustatory modalities. This review comprehensively introduces IMS-based processing systems (IMSPS), covering material innovations, device architectures, synaptic plasticity mechanisms, and applications across unimodal (visual, tactile, etc.), environmental, and multimodal sensing domains. Emphasis is placed on bioinspired robotic systems that leverage IMSPS for perception and interactions. Finally, we outline key opportunities in intelligent robotics while addressing challenges for IMSPS in scalability, neuromorphic sensing capability, and system-level optimization.
Keywords:
in-memory sensor
near-memory sensor
bionic robot perception
neuromorphic sensing
energy-efficient processing
Journal
M
IF:
22
Papers:
279
Citations:
0
Organization
Cited Papers