Recent advances in wireless and machine-to-machine technologies have shaped the evolution of embedded systems towards (mesh) networks of wireless devices incorporating multiple RF interfaces, sensors, actuators, data processing, decision making, and lightweight networking stacks. As the next decade will witness the interconnection of billions of networked embedded systems laying the foundations for connected automotive systems, smart cities, buildings, and grids, as well as attractive applications in the area of cyber-manufacturing, healthcare, and precision agriculture, it is essential to ensure that such systems are highly secure and that they can offer a resilient and dependable performance throughout their lifetime, so that they can autonomously and reliably fulfil their mission even in harsh environmental real-world conditions. SPIDR² builds on top of the ground-breaking results obtained within the SPiDR project on secure wireless protocols, ultra-wideband localization, context awareness, and tiny machine learning, towards the development of secure, resilient, and highly-performant solutions for wirelessly-networked embedded systems. Specifically, the project will deliver tangible and pioneering research results in the form of: (i) Secure and resilient localization systems based on ultra-wideband technology that can sustain a centimeter-level accuracy even in harsh RF environments. (ii) Innovative speech-based authentication and speech-to-text transcription services enabling an efficient speaker identification and the effective transmission of voice over low-bandwidth networks. (iii) Resilient and highly-performing protocols based on synchronous transmissions for reliable low-latency wireless communication.